“Superparameterization” is described by the Climate Process Team on Low-Latitude Cloud Feedbacks on Climate Sensitivity in an online meeting report (Bretherton, 2006) as:
a recently developed form of global modeling in which the parameterized moist physics in each grid column of an AGCM is replaced by a small cloud-resolving model (CRM). It holds the promise of much more realistic simulations of cloud fields associated with moist convection and turbulence.
Clouds have, of course, been the primary source of uncertainty in climate models since the 1970s. Some of the conclusions from cloud parameterization studies are quite startling.
The Climate Process Team on Low-Latitude Cloud Feedbacks on Climate Sensitivity reported that:
The world’s first superparameterization climate sensitivity results show strong negative cloud feedbacks driven by enhancement of boundary layer clouds in a warmer climate.
These strong negative cloud feedbacks resulted in a low climate sensitivity of only 0.41 K/(W m-2), described as being at the “low end” of traditional GCMS (i.e. around 1.5 deg C/doubled CO2.):
The CAM-SP shows strongly negative net cloud feedback in both the tropics and in the extratropics, resulting in a global climate sensitivity of only 0.41 K/(W m-2), at the low end of traditional AGCMs (e.g. Cess et al. 1996), but in accord with an analysis of 30-day SST/SST+2K climatologies from a global aquaplanet CRM run on the Earth Simulator (Miura et al. 2005). The conventional AGCMs differ greatly from each other but all have less negative net cloud forcings and correspondingly larger climate sensitivities than the superparameterization
They analyzed the generation of clouds in a few leading GCMs, finding that a GCM’s mean behavior can “reflect unanticipated and unphysical interactions between its component parameterizations”:
A diagnosis of the CAM3 SCM showed the cloud layer was maintained by a complex cycle with a few hour period in which different moist physics parameterizations take over at different times in ways unintended by their developers. A surprise was the unexpectedly large role of parameterized deep convection parameterization even though the cloud layer does not extend above 800 hPa. This emphasizes that an AGCM is a system whose mean behavior can reflect unanticipated and unphysical interactions between its component parameterizations.
Wyant et al (GRL 2006) reported some of these findings. Its abstract stated:
The model has weaker climate sensitivity than most GCMs, but comparable climate sensitivity to recent aqua-planet simulations of a global cloud-resolving model. The weak sensitivity is primarily due to an increase in low cloud fraction and liquid water in tropical regions of moderate subsidence as well as substantial increases in high-latitude cloud fraction.
They report the low end sensitivities noted in the workshop as follows:
We have performed similar +2 K perturbation experiments with CAM 3.0 with a semi-Lagrangian dynamical core, CAM 3.0 with an Eulerian dynamical core, and with the GFDL AM2.12b. These have λ’s of 0.41, 0.54, and 0.65 respectively; SP-CAM is about as sensitive or less sensitive than these GCMs. In fact, SPCAM has only slightly higher climate sensitivity than the least sensitive of the models presented in C89 (The C89 values are based on July simulations)…
The global annual mean changes in shortwave cloud forcing (SWCF) and longwave cloud forcing (LWCF) and net cloud forcing for SP-CAM are _1.94 W m_2, 0.17 W m_2, and _1.77 W m_2, respectively. The negative change in net cloud forcing increases G and makes λ smaller than it would be in the absence of cloud changes.
Wyant et al (GRL 2006) is not cited in IPCC AR4 chapter 8, though a companion study (Wyant et al Clim Dyn 2006) is, but only in the most general terms, no mention being made of low sensitivity being associated with superparameterization:
Recent analyses suggest that the response of boundary-layer clouds constitutes the largest contributor to the range of climate change cloud feedbacks among current GCMs (Bony and Dufresne, 2005; Webb et al., 2006; Wyant et al., 2006). It is due both to large discrepancies in the radiative response simulated by models in regions dominated by lowlevel cloud cover (Figure 8.15), and to the large areas of the globe covered by these regions…
the evaluation of simulated cloud fi elds is increasingly done in terms of cloud types and cloud optical properties (Klein and Jakob, 1999; Webb et al., 2001; Williams et al., 2003; Lin and Zhang, 2004; Weare, 2004; Zhang et al., 2005; Wyant et al., 2006).
(Bretherton 2006)
Dessler et al (GRL 2008) made no mention of strong negative cloud feedbacks under superparamterization, stating that sensitivity is “virtually guaranteed” to be at least several degrees C, unless “a strong, negative, and currently unknown feedback is discovered somewhere in our climate system”:
The existence of a strong and positive water-vapor feedback means that projected business-as-usual greenhouse gas emissions over the next century are virtually guaranteed to produce warming of several degrees Celsius. The only way that will not happen is if a strong, negative, and currently unknown feedback is discovered somewhere in our climate system.
There are a limited number of possibilities for such a possibility, but it is interesting that cloud super-parameterizations indicate a strong negative cloud feedback (contra the standard Soden and Held results.)
This is not an area that I’ve studied at length and I do not have personal views or opinions on the matters discussed in this thread.
References:
Bretherton, C.S., 2006. Low-Latitude Cloud Feedbacks on Climate Sensitivity. Available at: http://www.usclivar.org/Newsletter/VariationsV4N1/BrethertonCPT.pdf [Accessed June 12, 2009].
Wyant, M.C., Khairoutdinov, M. & Bretherton, C.S., 2006. Climate sensitivity and cloud response of a GCM with a superparameterization. Geophys. Res. Lett, 33, L06714. ftp://eos.atmos.washington.edu/pub/breth/papers/2006/SPGRL.pdf
Bretherton, C.S., 2006. Low-Latitude Cloud Feedbacks on Climate Sensitivity. Available at: http://www.usclivar.org/Newsletter/VariationsV4N1/BrethertonCPT.pdf [Accessed June 12, 2009].
Wyant, M.C., Khairoutdinov, M. & Bretherton, C.S., 2006. Climate sensitivity and cloud response of a GCM with a superparameterization. Geophys. Res. Lett, 33, L06714.
Climate Audit 
166 Comments
Cloud positive feedback is one of the most foolish and anti-common sense claims of the models.
This is particularly true of cumulus and cumulonimbus, which increase with the temperature during the day, move huge amounts of energy from the surface aloft, reflect huge amounts of energy to space, and fade away and disappear at night.
Glad to see somebody’s noticed, though.
w.
Re: Willis Eschenbach (#1),
I just returned from a two-week road trip through much of mid-USA. Most days were 90-100f, often 105+.
Tuesday (this week), it was slightly more bearable in Colorado at 4500+ feet as we drove west… only about 89 degrees. But still Very Hot. Our car was caked in baked-on bird poop; we were hot and tired. You get the idea.
In the distance, we saw a thunderstorm approaching. Suddenly, we were deluged with pelting rain, thunder, lightning. In less than two minutes, the temperature dropped from 89 to 66 degrees. The storm passed, but the cooling effect remained. Everywhere west of there was quite comfortable.
So, when (say) a 10km^2 storm drops surface temp from 89 to 66 in two minutes everywhere it goes (say… 20km/hr)… someone here can probably calculate the heat transfer rate away from the planet’s surface. Seems pretty significant.
For me, I was just impressed by this amazing natural air conditioner!
“unanticipated and unphysical interactions”: I know that people use “unanticipated” as a pompous alternative to “unexpected” or “surprising”; “unphysical” means, in the demotic, “garbage”.
So when its cloudy its not as hot as when its sunny (all other things being equal)Amazing!
Re: UK John (#5),
As one who has lived in warm climates, yes, it’s cooler under cloud than in the sun. But I have never been to the near Pole regions. Can the same be said of summer and winter at the Poles? Not a trick question, just hesitating to jump to conclusions.
Re: Geoff Sherrington (#12),
You have a very good point that should not be ignored.
When I first encountered the cloud data at ISCCP, I was startled to see the magnitude of the fluctuation in the total cloud amount during the past 25 years. It occurred to me at the time that the effect of the clouds could be season and latitude dependant (having lived in Alberta where winter clouds often accompanied warmer temperatures) so I plotted some UK data relating temperature and hours of sun (since that variable was available rather than cloud cover). The plots were by month with a fitted trend line:
The consistency of the seasonal slope direction was somewhat startling although the winter fits show higher variability. Whether the temperature level influences or is the result of the clouds (or both), it seems pretty reasonable that in winter the relationship is reversed here.
Re: RomanM (#14),
A class of model that catches my eye are “aquaplanet” models. From a mathematical point of view, it seems like a good idea to see what you can do without all the complications of continents and topography and introduce such things as tweaks on the base aquaplanet model. It also seems to keep people’s eye on the cloud issues a little more.
I looked at these at one time when I was interested in “symmetry breaking” – an interesting sort of-topological feature that arises even in aquaplanets. Symmetry breaking is interesting in MWP terms, since there is evidence (though not overwhelming evidence) that the ITCZ was further north in the MWP.
There is an interesting asymmetry in the actual earth with a relatively “warm” Arctic and cold Antarctic, but I think that it would be nice to approach this in the simplest possible variation from aquaplanet.
Re: Steve McIntyre (#15),
Variations between extremes of aquaplanet vs dryplanet would be a way of determining sensistivity to landmass distribution (say a hemispheric model: northern hemisphere land and southern ocean, or a two ocean model with simple land masses representing N-S americas and Asia-Africa). By not including topography and other geographic variables one might be able to determine how much of an effect gross land/ocean distribution has on climate.
Re: RomanM (#14), http://www.metoffice.gov.uk/climate/uk/about/UK_climate_trends.pdf
someone else had a go at this.
Re: UK John (#34),
Naw, they did different things with it. The closest they came to what I looked at was their correlation table for various variables on a seasonal and regional basis (table 15 on p.18). Although their results were based on 1961-2004 time period, they seemed to indicate that the relationship between temperatures and sun was strongest in the spring and fall, rather than summer which is what my graph seems to indicate.
Re: Andrew Dessler (#30),
I am always somewhat leery of arguments based on “we couldn’t find another reason for it”. These depend on the current knowledge of other factors and flaws in that knowledge can ovbviously invalidate the basis for that argument.
For example (and I am not offering this as a “theory” in any way, but as a “what if” – snip away if you wish, Steve), it appears to me that there is an assumption that climate sensitivity to forcings is a constant independant of the current climatic conditions. If that were not the case, a climate could be “kicked” by a catastrophic event (meteor,serious vlolcanic activity) into a state where the forcing takes much longer to return the climate to a different equilibrium. Unlike Sherlock Holmes, we can rarely eliminate all other alternatives to solve the mystery.
http://cdsweb.cern.ch/record/1180849?ln=en
Good. They have looked at clouds from both side now. Both positive and negative, but still some how, they haven’t realized they don’t know clouds at all.
This is fascinating. When climate change first appeared on my radar a bit less than three years ago, I initially adopted a very sceptical position, that warming due to CO2 was completely negligible. But then I moved to what I call ‘weak AGW’, that is, that there has been measurable warming due to CO2, but a fairly small amount, quite likely less than natural warming. One reason was this: that a simple model indicates a doubling would increase the temperature by a couple of degrees or so.
.
I think I’ve abandoned weak AGW for what I might call ‘extremely weak AGW’, which is roughly where I started. The ice core record clearly shows that CO2 followed temperature, and so could not be the driving force for the climate changes over hundreds of thousands of years. As far as I know, there is no proof that CO2 drove the climate in any period of the earth’s history.
.
But this creates a problem: how can the reality be so different to the simple physical model that predicts several degrees rise? Of course, it’s obvious the climate system is far, far more complex than the simple model. But I think the answer can be summarised with one phrase: negative feedback.
.
The fact that I’m sitting at my desk and writing this post strongly suggests that in the climate system negative feedbacks are the dominant factor. Of course, positive feedbacks can cause the occasional upset, but the system always returns to a reasonable equilibrium, and this is why the earth has supported life over the last billion years or so.
.
This work appears to be very important. If it enters the mainstream of climate science, as it probably should, then things are going to get very interesting! If these models indicate a significantly lower climate sensitivity then it will become obvious that much of the warming was natural, with enormous economic and political consequences. It may also help to restore the honesty of climate science. As long as those guys don’t use their improved GCM’s to predict the global temperature in a hundred years’ time….
@captdallas2 — I’ve looked at clouds from both sides now, from up and down and still somehow, it’s modelling illusions I recall. I really don’t know clouds at all
Perhaps I oversimplify, but I wonder why anyone is really surprised about cloud negative feedback. First, as mentioned above by Chris Wright, there must be considerable negative feedback in the climate systems or we would not be here.
Second, one would suspect that, if temperatures tend higher, more water evaporates and increases the opportunity for cloud formation.
Third, as can be observed on airline flights, cloud tops are very reflective, at least in the visible spectrum.
Given the huge amount of water on the earth’s surface, one would expect that cloud formation would be a major negative feedback mechanism.
But, again, perhaps I oversimplify.
UK John,
I live way up in the Swiss mountains. During the day, it’s ok what you say, but at night it is the other way around. Clear night equals cold, overcast nights means much warmer nights.
Fred
Hey Roy Spencer–you were right!!!
But there is a simple counter to these results–ignore them.
Roy Spencer blogged about a newer report by Caldwell and Bretherton here:
http://www.drroyspencer.com/2009/01/new-study-doesn%E2%80%99t-support-climate-models-but-you%E2%80%99ll-never-hear-about-it/
He was right – everyone ignored it.
What has amazed me is the fact that the importance of clouds has been known since at least the 1850s. Arrhenius included accounting for clouds in his analyses in 1896. The numerical value of the innocent-looking Earth’s Albedo, one of the foundations of the simple zeroth-order radiative-equilibrium model, is due primarily to clouds.
In engineering, a phenomena or process known to be a controlling dominate factor would receive all the investigations needed in order to pin down exactly what’s happening. It seems that the climate change community to this day hasn’t been bothered to determine the facts with respect to clouds.
The present situation in the case of clouds is yet another data point which clearly indicates that the critically important aspects of the Earth’s climate systems are not handled ‘based on first principles’ as is so often advertised by those who have circled their wagons. The US EPA recently informed me again that CGMs are based on the fundamental laws of physical phenomena and processes. I predict that as resolution in the discrete grids is increased and parameterizations become more nearly based on fundamental principles, there are lots of additional surprises in store.
Here is the abstract of a review article on clouds from 2005. The full article is behind a paywall.
Steve: I know the point that you’re making, but don’t go a bridge too far. One can dispute the relative priority of cloud research but it’s not as though the issue has been totally ignored. The present study is, after all, resulting from an official program.
Caldwell and Bretherton 2009 is available here. In their opening paragraph, they say:
People have to be careful about being overly friendly to an article because they “like” the results. I don’t know enough about the issues relating to this matter to have an opinion – I am merely putting some apparently interesting issues up for discussion.
In a practical sense, given that clouds have been the big uncertainty since at least the Charney Report, you’d think that this would constitute about 50% of the research budget, but it seems to be a pretty small field.
The Stephen’s review article on clouds in the Journal of Climate is freely available here.
Certainly worthy of more study.
So, clouds warm when it’s cold and cool when it warm. The exact action one would expect from a balanced system. Low and behold this agrees with observation of billions of years that our climate system is indeed balanced within a few degrees.
Re:#’s 4, 10, and 14
Strictly as a lay lurker, it has been my experience that the threshold between the clouds making local weather cooler or warmer is a temperature somewhere below say 40 degrees F, depending on many factors. My hunch is that the majority of the surface area of the world is on average hotter than 40F, therefore the net global cloud feedback will be negative at our current climate state, and get more negative if we continue to warm.
#1 Willis
Yes! No one has ever explained to me why this simple observation should not also be the key to cloud feedback on much longer timescales as well.
I also found this paper as well:
Wyant, M.C., Khairoutdinov, M. & Bretherton, C.S., 2006. Climate sensitivity and cloud response of a GCM with a superparameterization. Geophys. Res. Lett, 33, L06714.
Steve: This was discussed in the above thread.
Fascinating! Definitely worth looking into in more depth…
I see it was already furnished. Sorry for that. I guess I need new glasses…
After seeing a cloud data based albedo reconstruction a year or two ago, some things became apparent. Albedo, even as described in K&T(97?) breaks down the total, about 0.30 into atmospheric contributions of 0.22 and surface of 0.08 with around 60% cloud cover. The reconstruction (Palle’ & Goode 07) shows variations in albedo, peak to valley, of around 10% back in 1998. Using K&T type average numbers, one sees that there’s about 342 W/m^2 average incoming solar at 1 AU and around 105W/m^2 of reflected (albedo) power, leaving around 240 w/m^2 absorbed power that must be radiated by the Earth. A 10% variation in albedo from cloud cover change amounts to around a 10.5w/m^2 short term change in the actual absorbed power. Short term here is not quite seasonal short, but close.
The 10.5 w/m^2 change occurred around 1998, probably due to ENSO El Nino activity. Since albedo is poorly known and has been assumed to be a constant or slowly varying, it is an achilles’ heal for the vast majority of sensitivity measurements. The fact it can vary that much on relatively short time frames means it cannot be left as some averaged constant and it is not varying but it is changing substantially more than any other factor that might be used in a sensitivity measurement. As for actual measurements, there’s scarcely 30 years of data that I’ve found and it has serious gaps.
One thing I haven’t seen done by anyone (else) is the construction of a cloud cover fraction versus incoming and outgoing power. I did a back of the envelope very crude one last year when I had a bit of time to spend. Two points pretty much describes what is roughly a straight line. Outgoing, one has clear skies which is the usual surface radiation partially blocked by ghgs on one end and overcast skies on the other, which is close to 100% blockage outbound at the cloud plus BB emissions based on cloud top T. For the incoming solar power, clear skies pretty much gives the whole 341 W/m^2 minus the 0.08 albedo so there’s lots of excess incoming power over emitted power. Cloudy skies blocks nearly 90% of the solar so very little is entering. The two lines intersect at around 60% cloud cover (surprise surprise) with less cloud cover yielding a net incoming power while more cloud cover yields a net outgoing power. Radiative transfer doesn’t define why there is or should be 60% average cloud cover or why it might vary as that is in another domain. Also, using some sort of average cloud is quite different from the multitude of cloud types, locations and actual effects but it does suggest a negative feedback mechanism and makes intuitive sense as part of a real system where the added thermal power coming in strikes the surface, evaporates more moisture, … water vapor cycle stuff … which forms more clouds, creating the negative feedback T regulation. Of course, once one has clouds and cloud formation going into the mix, the prospects of cosmic ray influences along with every other particulate pollution, natural or otherwise, starts popping into possibility.
Clouds also force out the simple averaging type modeling as many types like to form during the day, etc.
It’s really good to see that someone is actually trying to take the stuff into account with the video gamers.
Steve: For editorial reasons, no further mention of cosmic rays on this thread.
Worth re-reading in this contextg is the RC discussion in Aug 2006 of Soden and Held 2006 a standard survey of feedbacks. The main thread states:
#16 Held says:
Clouds? negative feedback? …harrumph… settled… robust… heavy sigh, sneer & eye roll… consensus… aerosols… Secret Warming Pipeline larger and deeper than we thought… just means warming will be worse when it comes… closing additional heavy sigh and harrumph.. then …official silence.
The reason I tend discount the existence of a large negative cloud feedback is that it appears to be inconsistent with the paleoclimate record. We see large swings in temperature over the past 50 million years (at least), and a strong net negative climate feedback, which would stabilize the climate, should not allow the climate to oscillate like that. Jerry North told me the other day that he spent several years back in the 1980s trying to explain the ice age with present-day CO2 levels, and could not, suggesting that CO2 plays a key role in the ice ages. With a net negative climate feedback, I don’t see any explanation for the ice ages — or for climate swings further back in time.
Re: Andrew Dessler (#30),
Thanks for the comment. I wish IPCC would spend more time on questions like this, as they are the sort of thing that scientists from other disciplines and the educated public wonder about right away. What would you recommend as a reading on this topic?
While there’s considerable variability in ice age terms, equally puzzling is the relative non-variability in “deep time”. That the planet has not veered into permanent Icehouse or permanent Hothouse indicates an underlying stability that seems to be opposed to strong positive feedback – though I understand where you’re coming from.
As a sidebar here, there’s an asterisk for me attached to some of the early Icehouse periods. Some Canadian geologists do not interpret the tillites as evidence of past glaciation (rather as turbidites or something like that.) Early Icehouse is a pretty recent theory and might fall apart on geological grounds. I don’t have a firm opinion on this, but the geological critique is non-trivial.
Re: Andrew Dessler (#30),
The problem with this argument is that it at least borders on the false dilemma logical fallacy and possibly begging the question as well.
Re: Andrew Dessler (#30),
http://en.wikipedia.org/wiki/Argumentum_ad_ignorantiam
Steve: TRy not to throw slogans around. I know what Dessler meant here. Whether it’s right or wrong, the point isn’t trivial.
Re: Andrew Dessler (#30),
The problem is that we know that ice-ages are driven by insolation, not primarily CO2. The argument is that with high climate sensitivity the insolation changes (somehow) modulate the amount of CO2 which in turn drives climate. However this ignores the fact that throughout most of Earth history this clearly didn’t happen. The Milankovic signal is there all the time (it is visible in sediments under favorable circumstances), but most of the time the climatic effect is small. Does climate sensitivity go up as CO2 level goes down?
Also it is hard to see how the CO2 theory accounts for the shift in glacial cycles from the 41-kyr obliquity cycle to 100-kyr eccentricity cycle 800,000 years ago.
Steve: While I agree that the evidence for the very early glaciations like Pongola and the Huronian may be shaky, I have seen deposits from the Late Proterozoic and Late Paleozoic glaciations in several areas on different continents, and they seem quite certain to me. Whether the Late Proterozoic glaciations ever really went to a complete “Snowball Earth” is more uncertain.
Re: Andrew Dessler (#30),
If a system has strong negative feedback it will not guarantee a smooth and constant output. A poorly-designed servomechanism, although based on negative feedback, may oscillate badly and respond to external step changes with oscillations that may take some time to die down. As an electronics design engineer in the semiconductor industry for thirty years I have had a lot of experience with misbehaving servomechanisms. Looking at climate graphs (and stock market graphs) I’m often reminded of marginally stable negative feedback systems responding to external shocks. In the case of climate, external shocks might come in the form of big asteroid impacts, volcanic eruptions, extreme solar variability and changes in the orbit. These effects may cause extreme climatic changes such as Snowball Earth, but eventually negative feedback brings the climate back to a reasonable mean.
Dan Hughes:
The DOE established a very large program to study clouds quite some time ago. You can read about DOE ARM at any of these google results. Cloud research is also funded by other programs.
While this is an improvement, we still have to remember that we are talking about the GCM models. They have been falsely geared to a high sensitivity by assuming that all the warming of the late 20th century was caused by human emissions of greenhouse gases; an assumption that is obviously wrong for many reasons. ‘Adding in’ the negative feedbacks from the improved cloud model will result in a lower sensitivity, but one that is still too high, due to the erroneously high ‘starting’ sensitivity in the models.
-snip.
Steve: Please do not over-editorialize. While you may disagree with people, please keep in mind that many serious people hold their views in good faith.
Dear All
Just to say that my take on the possibility of low sensitivity is similar to Andrew Dessler’s…that is, you can’t explain the palaeo record without reasonably high sensitivity. If the models say one thing, but our field/palaeo evidence says another, I go with the evidence. I remember saying roughly the same thing on CA some months ago when we were discussing Ross’ work on GDP as a climate predictor I think. Steve: you kindly asked if I’d write a post on this but I really haven’t had the time (day job!).
Re: san quintin (#36), In the absence of the requested post :), can you suggest any relevant readings on the matter?
Given that you and Andrew Dessler (and for that matter, Hansen at last year’s AGU), place such weight on this particular line of argument, it’s furstrating that it isn’t addressed more prominently (or seemingly at all) by IPCC.
Also I’ve got the same question for you as I had for Andrew Dessler – it’s hard to find a line that yields high sensitivity for Ice Age swings without also veering off into a permanent Icehouse or Hothouse, an issue that isn’t always faced very squarely.
Oh boy, the big climate changes aren’t allowed by negative feedbacks argument again. Because of course we understand how a tiny change in insolation can cause ice ages anyway. Right. Of course, the change in insolation varies with latitude, season, etc. but it is still okay to treat it as a homogeneous forcing. Right.
Okay. So paleoclimate is the basis for positive feedback arguments now. I’m sure that those here who have long noted problems with paleoclimate studies will have some comments about this…
Steve: calm down…
Re: Andrew (#37), Sorry, haven’t eaten lunch yet…
Folks, please do not debate the “CO2 theory” in every thread on every occasion. And before people pile on, I’m more interesting in seeing what references support this particular paleoclimate argument.
Re: Steve McIntyre (#43), As no stranger to the argument, I’ve come across some references. This is an early one:
Hoffert M.I., and C. Covey (1992) Deriving global climate sensitivity from paleoclimate reconstructions. Nature 360:573-576
Hi Steve
I think that Annan and Hargreaves 2006 paper “Using multiple observationally-based constraints to
estimate climate sensitivity” in GRL is a good start. Re: tty (#42), Don’t forget that there are also geological controls which probably initiated Quaternary glaciations. After this, insolation plays a role (mainly in northern latitudes) amplified by CO2. The point is, if the climate were insensitive to changes in the forcings, then it is highly unlikely you could generate Quaternary ice sheets. Similarly, a favouraite sceptics position is to argue for a global MWP and LIA…both of which (if true) also suggest strong temperature responses at a global scale to relatively minor forcings. It seems improbable that you could have a global MWP AND low sensitivity.
Re: san quintin (#45),
Considering that we don’t know the magnitude or even the origin of the forcing(s) that caused the LIA and the Dark Ages cold period, it’s rather disingenuous to imply that they were small. Speculation about the Maunder Minimum and decreased TSI as the cause of the LIA is just that, speculation. We have sun spot counts and 14C data and that’s about it. One can easily argue that current conditions are similar to the Roman and Medieval warm periods and that some still unknown factor or combination of factors caused temperatures to decrease during the Dark Ages and the LIA.
Re: san quintin (#45),
snip – personal opinion. I’ve asked people not to express opinions – find a citation for your view.
snip
Re: Andrew (#46), I’m afraid I don’t understand your point. I wasn’t being snide (or at least I didn’t think I was).
The sensitivity numbers from the “superparameterized” models appear to agree with Richard Lindzen’s sensitivity, 0.5, derived from an empirical approach. I am a layman, but if this is true wouldn’t it be significant?
General Comment
Now that a different cloud parameterization has predicted negative feedback instead of positive feedback, everyone seems to forget there are very basic mathematical and numerical problems with all of these numerical models. I repeat again that a numerical model is not a mathematical proof of anything and can only be useful if it accurately approximates the unforced continuum differential equations (recall that the IBVP for the hydrostatic differential equations is ill posed so that any numerical method will fail to converge to the continuum solution and the nonhydrostatic equations contain fast exponential growth components that cannot be accurately approximated by any numerical method). In the previous sentence I am referring only to the unforced PDE’s (so called dynamics) and have not even mentioned problems with inaccurate approximations of the physical processes ( the parameterizations).
What this new “result” shows is that the models can produce anhthing one wants by appropriately tuning the forcing, exactly as I have stated (and proved) in previousthreads on this site.
Has anyone discussed how ludicrous it is to have a cloud parameterization dependent only on a vertical column of model data from a poorly resolved numerical approximation of an ill posed system of equations? Or what the impact of a superparameterization means when it is based on such model data.
Jerry
Re: Gerald Browning (#48),
I do not think anyone has forgotten the issues you raised earlier about GCMs. Instead of being unhappy with this thread, I would have thought you would be pleased these papers show you can get different predictions based on the hypothesis the GCM is modeling. In fact, this thread reminds me of the “Ryan’s Tiles” thread where it was asked “What kind of Antarctica do you want?”
I would still like to see the results of a GCM that could model the climate regimes such as we have seen in 1910, 1945, 1975 and 2007.
Why does the sensitivity have to be symmetric WRT temperature? As temperatures rise above freezing, more water vapor is available for cloud formation. Clouds are purported to be the moderating influence on heat build-up. But going the other way, once the temperature is below freezing, cloud formation would cease. This would allow in all incoming radiation, but then the resulting ice and snow would reflect much of it. I guess you could say the sensitivity could itself vary with temperature. I guess the devil is in the details in the moderate temperature region. Also, the moderating mechanism could fail with higher solar input – chaos could literally ensue.
Re: Jim (#50), Hi Jim. yes, you are right that sensitivity is not an unmovable figure (partly why there is uncertainty). It varies with temperature, vegetation, ice sheet extent etc. But estimates from recent events (Maunder Minimum) and Pinatubo also suggest a sensitivity higher than 1-1.5C.
snip – I asked that people not debate CO2 on every thread,
Re: Andrew (#51), OK. Perhaps I should have said simply “Arguments for a global MWP and LIA also require strong temperature responses at a global scale to relatively minor forcings. It seems improbable that you could have a global MWP AND low sensitivity”.
Re: san quintin (#52), There is some evidence for strong increases in cosmogenic isotopes at the MWP and low levels at the LIA, which are not “small changes in forcing”.
snip – I’m not interested in this thread being diverted to the discussion of speculative theories.
This post is entirely correct in flagging up the importance of clouds but I contend that clouds are just one part of a broader negative feedback process controlled by rapid changes in the global air circulation both geographically and in terms of intensity as a result of any extra warming of the air alone (as opposed to extra warming of air and ocean).
Changing ocean surface temperatures induce air circulation changes as the air seeks to restore the sea surface/surface air temperature equilibrium and at the same time resolve ocean induced variations in the sun to sea / air to space equilibrium.
The air circulation changes alter all the processes involved in the rate of energy transfer from surface to space. Cloudiness changes are an inevitable by product of changes in the air circulation systems.
In due course stabilty is always restored between the four said parameters (sea surface / surface air and sun to sea / air to space).
Only huge catastrophic changes capable of altering the temperature of the whole body of the oceans can set a new global equilibrium in the short term (less than millennia). The sun can also do it gradually but it takes centuries e.g. from Roman Warm Period to Mediaeval Warm Period to Little Ice Age to now. The solar effect is heavily modulated over time by ocean cycles. A change in the composition of the air alone cannot do it due to the thermal inertia of the oceans combined with the speed of responses available in the air.
The role of water vapour combined with the latent heats of evaporation and condensation gives the air circulation changes the major part of their ability to accelerate energy transfer from surface to space.
So, the most common and by far the largest forcing at any given time is multi decadal variations in energy emissions from the oceans. In the background are slow century scale changes in solar output.
Temperature changes induced by sun and oceans drive air circulation changes which drive changes in every aspect of climate including convection, conduction, evaporation, condensation, precipitation, windiness, cloudiness, albedo and humidity as regards both quantities and distribution.
Water vapour in itself is not a driver nor does it have cycles or periodicities of it’s own. It’s a very useful contibutor to the whole process though and without it the Earth would be entirely different
Any ocean surface warming is caused by solar energy previously absorbed working it’s way back to the surface. It is becoming clear that oceanic energy emission to the air is not stable on multidecadal time scales.
The air circulation has to balance both the energy flow from ocean to air with the energy flowfrom air to space AND the energy flow from sun to ocean with the energy flow from air to space over time. Everything we observe is a feature of that interplay.
To deal adequately with any warming of the air from extra CO2 or any other increased GHG the air circulation and weather systems just shift their size and/or positions to adjust the rate of energy emission to space to restore equilibrium between air and oceans.
It is the latitudinal position of the weather systems which is most significant in changing the rate of energy flow from surface to space. Secondary to that is the speed of the hydrological cycle.
The equilibrium which the weather systems work back towards is set by the rate of energy flow from the sun modulated by the rate of energy flow through the oceans. It is not set by the characteristics of the air which is the point at which I come to the conclusion that the work of Tyndall and others is faulty.
The air circulation changes ensure that over time the energy radiated to space matches the energy received from the sun despite disruptions in the flow caused by the effects of the ocean cycles or changes in the composition of the air.
Everything we see in the air and the oceans is part of that natural energy balancing interaction and human emissions have no part to play other than a very small insignificant human induced shift in positions or intensities of the main air circulation systems. Wholly imperceptible in the face of natural variability.
Ramanathan has some thoughts on this subject (and albedo)
He also comments quite succinctly.
Re: Andrew Dessler (#30),
80% of albedo is atmospheric, substantially cloud, at least at present time. Even if there is a serious net negative feedback from clouds in operation at present time, glaciation over substantially lower lattitudes than currently receive winter snow cover is going to short circuit any cloud feedback during these glaciation periods. Even the absence of clouds blocking incoming insolation is virtually going to have no little to no effect (on incoming power) if the surface has a significant albedo as well.
Also, if internal oscillations like the 1998 ENSO can affect albedo by 10% as suggested by the Palle & Goode 07 paper, one must take the concept of just what strong feedback really means in this case. Cloud formation is what this proposed feedback is about. There are many types of clouds and many conditions required for their production. What’s more there are external sources that affect formation and some of those are random and ENSO is evidently chaotic and has substantial effect upon cloud cover, and hence, an effect upon albedo in nonglaciation periods.
I’ll foregore all the various impulse possibilities already mentioned above along with orbital cycles previously mentioned as I suspect all probably contribute to some extent or another. Except to say I thought it was more Nir Shaviv who was into galactic orbit bobblings in and out of the spiral arm and that Svensmark was more into other related factors.
Steve-
I would recommend looking at this paper: The equilibrium sensitivity of the Earth’s temperature to radiation changes, Hegerl and Knutti, Nature Geoscience 1, 735-743 (26 October 2008) doi:10.1038/ngeo337. This is a pretty good review paper that contains a lot of good references on this problem.
I guess I have not thought about why variations of the deep time climate are not more variable. One reason is that it seems quite difficult for the planet to get out of a snowball Earth configuration. I don’t think anyone has a good theory for that.
Roman-
I think the argument against a negative feedback is more nuanced than that. We have a theory of a CO2-driven net-positive-feedback climate that explains most of what we see in over at least the last 50 million years: the Eocene (high CO2), ice ages (low CO2), the modern warming (increasing CO2). We may not be able to explain 100% of the data, but we can explain a lot.
If we invoke a negative feedback, then we suddenly can’t explain anything: why was the Eocene so hot, why do ice ages occur, why is the climate warming now? There may be legitimate explanations for everything that does not rely on CO2, but I have not heard it.
Thanks!
Re: Andrew Dessler (#61),
If anyone finds a copy of this paper online, would they please post the link?
And thanks to Andrew Dessler, san quintin, Steve and others for a most interesting new discussion!
TIA,
Pete Tillman
Re: Andrew Dessler (#61),
Perhaps it is a function of explanatory space limitations, but there is something not very satisfying about these nuanced conjectures. Important to note in the above theory would be what initiates the increase/decrease in CO2 levels in the atmosphere. A few words on the separation of feedbacks from surface albedo, clouds and atmospheric moisture might be helpful. A word or two on the cause/effect of increasing/decreasing temperature on increasing/decreasing atmospheric levels of CO2 would help. At what point, if any, would methane levels in the atmosphere become important. And as noted in a DeWitt Payne post on this thread can we assume that the feedbacks effects (particularly clouds) are always the same or can they be a function of temperature and other conditions unique to a given climate era?
Re: Andrew Dessler (#61),
Re: Peter D. Tillman (#63),
“The equilibrium sensitivity of the Earth’s temperature to radiation changes”, Hegerl and Knutti, Nature Geoscience 1, 735-743 (26 October 2008) doi:10.1038/ngeo337.
Online at http://www.iac.ethz.ch/people/knuttir/papers/knutti08natgeo.pdf
Steve, thanks for the link. This is an informative review and a good way to get up to speed on the climate-sensitivity topic — Andrew, thanks for suggesting it. As Steve points out, H&K do seem a little short on low-sensitivity viewpoints, but they lay out multiple lines of evidence that climate sensitivity is probably around 3 deg. C for 2xCO2. I’m not entirely convinced.
Their Figure 3 is the heart of the review. Someone with better web skills than me might extract this page from the pdf and post it, so we can discuss it more easily. This would qualify as Fair Use (use for critical commentary) and so wouldn’t violate the authors’/journal’s copyright.
TIA,
Pete Tillman
Re: Peter D. Tillman (#111),
“Their Figure 3 is the heart of the review. Someone with better web skills than me might extract this page from the pdf and post it, so we can discuss it more easily. This would qualify as Fair Use (use for critical commentary) and so wouldn’t violate the authors’/journal’s copyright.”
Extraction was the easy part.
I have no idea how to post it here – the extracted JPG is about 1.2Mb at a readable resolution
This has the potential to become a very interesting thread – I would love to see a sharp debate between Dessler and Spencer, but I’m somewhat fearful that Steve Mc’s narrowly-based snipping may discourage them. A real debate such as this has not been done in public to my limited knowledge.
Re: ianl (#117),
To post it you can put it on your own website or you need to get an account on an image hosting site such as flickr, imageshack, or tinypic (you can’t put it directly into a comment). Post it there and then take that URL and use the Img quicktag and paste the URL into it. Or you can use the Link tag which will add a hyperlink to the image instead of displaying it inline.
Re: ianl (#117),
Ianl, You can register on my site and post your image in the Data Repository. You’ll then be able to link to it from anywhere. http://whatcatastrophe.com Let me know there if you need assistance.
Re: Jeff Alberts (#119),
Nice offer, Jeff, but after 6 tries at the “Catcher in the Rye” I gave up
Re: Peter D. Tillman (#111),
Concerning the recommended Knutti and Hegerl paper .
It is indeed a good entrance in the sensitivity problem as it is usually presented exactly in this manner .
That’s why I will limit the analysis to this particular paper and the criticism applies mutatis mutandis to every other paper discussing climate sensitivities on the “GCM mode” and this mode is 90 % of the papers .
1)
The introduction states the topic :
The main criticism is already with this opening statement .
The foundation is here that the system is in equilibrium and that deviations from equilibrium can be treated by a perturbation theory .
The first assumption is trivially wrong . The system is NOT in equilibrium , has never been and will never be .
The day half receives much more energy than it emits and the night half emits much while receiving nothing . The sum of both is not zero .
Additionaly the radiative unbalance is not distributed homogeneously . The presence or absence of clouds modifies the unbalance significantly and is able to change the temperatures by 10 °C or more in either directions .
There is not a pseudo equilibrium over long time scales either as is demonstrated by the variation of temperatures averaged over such scales .
So the paper already begins with a non defined and non existent concept – the system in equilibrium .
The perturbation theory is generally a first order linearisation around an equilibrium point . We have already seen that the Earth system is NOT at the equilibrium point but even if it were , perturbation theory would not work .
Why ?
Because the perturbation theory can only work if the system didn’t evolve much from its equilibrium point .
The system atmosphere + oceans being chaotic at short scales (weather) , on multidecadal scales (ENSO) and on still larger scales (Earth orbit) , it diverges from any initial state very fast . Perturbation theory applied on strongly non linear equations (like Navier Stokes f.ex) what is the case for our Earth system cannot give results that converge to the true behaviour .
In a way this is the same thing that Jerry Browning has already shown here several times – the GCM are either unphysical (hydrostatic assumptions) or exponentially divergent .
Still another way to make understand this issue – making the equilibrium assumption in a study of Rayleigh Benard flow (behaviour of a fluid heated from above) would consist to … cut off the heating .
But if I cut off the heating , the Rayleigh Benard flow disappears because the non equilibrium conditions are necessary to observe it . So whatever conclusions I could draw from an equilibrium assumption and perturbation treatment would be irrelevant and non observable .
.
2)
There is no physics inside . There are actually no mathematics either with the exception of a small bit of statistics at the end (skewness and PDF) .
The authors admit themselves that the range of the “sensitivity” is the same in 2009 as it was at Arrhenius time 120 years ago . Arrhenius is actually mentionned in the references !
How is that possible when Arrhenius ignored everything about non linear dynamics , quantum mechanics and matter/radiation interactions ?
Well it is because the simple equilibrium physics at his time was still the same as today .
That is also why it is irrelevant how many numerical climate models exist or will exist – they all share the same equilibrium assumptions and can’t therefore contribute in any way to make progress in the “sensitivity” issue .
What makes matters worse is that the paper considers climate models as an alternative reality so that running a model is equivalent to do an experience in a lab .
It reaches summits in this statement :
.
Of course other much more realistic ways exist .
Notably it would be about time to take the Nature seriously and to treat it as a dynamical system what it is instead of droning all the time about equilibrium temperatures and time independent states .
However dynamical systems are notoriously difficult to study and very unfriendly to numerical models (D.Hughes and G.Browning explained x times why it is so) .
That’s why it is to be expected that people who invested billions in equilibrium numerical models will NOT switch easily to methods where these models are just useless balast .
But that is also why it is to be expected that a breakthrough in the understanding of the climate will come OUTSIDE of the GCM community because it is only there that the dynamical approach using modern physics is privileged instead of numerical simulations based on non realistic assumptions .
An example of such approach is given here :
.
A special note for Steve . In the linked paper you will see what is the relationship between the trajectories in the phase space of a chaotic system and the eigenvectors of an auto covariance matrix .
More specifically it is proven that one can truncate the number of eigenvectors to the number which is equal to the dimension of the chaotic attractor .
Re: TomVonk (#126),
Thanks TomVonk – very clear and well written.
Re: TomVonk (#126), Tom, why don’t you publish your thoughts quantitatively? You and Jerry Browning together could write a terrific paper.
This is a most interesting and aposite thread, as it goes to the heart of the debate.
Yesterday was the coldest day in 43 years in Canberra (and the coldest on record where I live in the Tuggeranong Valley). It is winter we are at latitude 35S, 2000ft and inland. Thursday night was clear and frosty, but clouds moved in early in the morning, so we had minus 2 to plus 3. On Friday night it remained cloudy, so we didn’t get a frost, and as I write this Saturday lunchtime we have clear fine weather and a nice hot (over 10 degrees) day.
I conclude from this that cloud acts as an insulator, impeding the heat flow. At night, when the surface is hotter than outer space, the heat flow is impeded by cloud, so the surface temperature will be warmer (less frost or no frost) when it is cloudy. During the day, when the sun is heating the surface, a cloudy day will be colder.
My completely unsophisticated simple model is for a 30% land (no evaporation) 70% water (evaporation) planet and is based on Stephan-Boltzmann and Clausius-Clapeyron. I get a sensitivity of 0.1 DegC/W/m2 for the SURFACE. Interestingly the difference between Winter and Summer average temperatures, and between day and night temperatures in Canberra imply a sensitivity of this order, not of a much higher sensitivity.
Ou in a paper that I discussed a couple of years ago here considered the question of why the earth’s climate had been “relatively stable” (i.e. neither Hothouse nor ICehouse) for hundreds of millions of years. Re-reading the post, it’s interesting that Ou hypothesized low clouds as a key stabilization mechanism.
Re: Steve McIntyre (#64),
Thanks for the link to that 2006 post. Has it really been three years I’ve been repeating this bit of doggerel?
I think I’ve never heard so loud
The quiet message in a cloud.
===================
Folks, if your comment doesn’t tie in to cloud sensitivity somehow, please save it for another occasion.
And PLEASE no cosmic rays or personal theories of how things might have happened.
#61. Andrew, given the importance that you assign to this line of argument, I don’t recall seeing it presented in IPCC TAR or AR4. Is there a relevant assessment in the two reports that I’ve overlooked?
Here’s a relatively recent reference on clouds and climate sensitivity published in Science 4/11/2008: Amplification of Cretaceous Warmth by Biological Cloud Feedbacks
Re: DeWitt Payne (#66),
The implication of the paper I linked above is that cloud feedback is a function of the temperature regime. If the authors’ theory is correct, cloud feedback is positive at the temperature of the Eocene Optimum. However, there could still be negative feedback from low clouds at current temperatures.
Steve, have you taken a look at Ole Humlum’s site http://www.climate4you.com ? The climate and clouds page has a lot of info about cloud forcing including negative feedback. The cloud forcing (a 5% reduction since satellites) is a much larger forcing than CO2 and seems to have been missed by most of the climate community.
Cheers
Cloud composition, particularly low cloud, is not fixed: ie see
http://www.newscientist.com/article/mg20126903.800-groundbased-bacteria-may-be-making-it-rain.html
and Salter and Latham:
“The amount of solar energy reflected from low level marine stratocumulus clouds at present is a
function of
· the liquid water content,
· the size of the drops,
· the time of day,
· the latitude,
· the amount intercepted by high cloud,
· the depth of cloud,
· the reflection coefficient of water drops.”
If these effects are dependent on temperature change then one ends up with a situation where the sensitivities — CO2, sea surface, cloud formation, cloud droplet coalescence — are all variables. It is surely too early to claim that we know CO2 sensitivity to any degree of accuracy.
JF
But I’d still bet on .6.
Let’s take them separately, because they’re different kinds of events.
Maunder was a low frequency event, and probably relates to sensitivity from our current condition down. I don’t think we have enough data to model it properly, but I can’t say for certain. I will note that down is not up. There is a hazard in trying to reduce sensitivity to some linear expression.
Pinatubo was an impulse event. I really don’t see a sound basis for trying to determine the low frequency characteristics of a system based on an impulse response. I would be hesitant to even try, as the results are too likely to be deceptive.
As an example, in electronics, capacitors and inductors have a “self resonant frequency”, and their behavior is opposite. That is to say, above the SRF, capacitors act like inductors and inductors act like capacitors. If you’re trying to figure out what kind of part you have in your black box, and you only have high frequency data, you’re very likely to get it backwards.
Tying this all back in to clouds…
The vapor pressure of water is exponential with temperature. The growth of cloud droplets is a function of water vapor density. This points to some significant non-linearities in cloud formation. Why would you expect this to result in some linear sensitivity?
Re: Dishman (#69),
Dessler: “…But estimates from recent events (Maunder Minimum) and Pinatubo also suggest a sensitivity higher than 1-1.5C.”
Dishman: “Pinatubo was an impulse event. I really don’t see a sound basis for trying to determine the low frequency characteristics of a system based on an impulse response. I would be hesitant to even try, as the results are too likely to be deceptive.”
The Pinatubo natural experiment was discussed in some detail at http://www.climateaudit.org/?p=1335, Water Vapor Feedback, starting around #84. Soden et al. got a sensitivity of around 1.7 – 2.0 deg. C, and Douglass & Knox, a lower number. Douglass & Knox also found a short response time and negative feedback to the volcanic forcing. D&K:
Their results were criticized at the time (2005-2006) — see the Wigley, Ammann comment in that thread, #87.
Both of these Pinatubo sensitivities are thought to be lower bounds, iirc.
Peter D. Tillman
I think you’re making a mistake leaving CO2 out of the discussion. Andrew Dessler brought it into the discussion as a reason for ignoring those model results (mentioned above) and presumably the observational results by Spencer. He is not a minor player; his work is being touted as final proof of net positive water vapour feedback yet he indicated that ignoring potential negative feedback is largely a gut feeling because with CO2 we can explain a higher percentage of the paleo record. If the CO2 argument is fundamental in the clouds discussion for him then so it should be for the rest of us.
-snip
Steve: of course it’s the issue of interest. Unfortunately every discussion of the “big” issue covers the same points and has the same sort of piling on. Editorially I prefer to stick to sub-issues in the hope that some footing can be achieved on smaller points. There are many other places where you can exchange opinions on the “big” issue.
Similarly I’m uninterested in presentation of “fringe” theories. At this site, stick to discussion of conventional theory.
Re: JamesG (#72), Much more important point about the glaciation cycle can be made that presumably doesn’t break the rules….That glaciations are initiated by Milankovitch forcings which, while resulting in net radiation changes so small as to make it impossible to explain how they could have actually happened even with positive feedback, are not in the least comparable to spatially and temporally homogeneous forcings like CO2. Some papers have been written about the importance of this point (for instance). Further back things get fuzzier and I think it is dangerous to just declare any effect you see in the geological record as due to some tiny forcing. God only nows what was in play during the Eocene, and it’s pure hubris to claim knowledge is that case. The Maunder is…tricky. First, no doubt we are only talking TSI. Well, okay, except that TSI proxies are all over the place, as are climate reconstructions (we would know, eh?). And 20th century…again forcing estimates are a huge wash, plus there may be important factors in play which we don’t know about. As to Volcanoes, many papers have been publish on this subject including on Pinatubo and to suggest that positive feedback is required to explain it is, well, just wrong.
The paleo argument is contingent on many assumptions which I am not impressed by. But it is popular. Go figure.
Steve: please stop expressing editorial opinions. Discussing a published paper is excellent, but try to be descriptive and leave your “impressions” out.
Interesting post. From what I have read on climate models the effect of clouds is one of the most difficult to model. I think if people also step back a bit you can see that the reason why a lot of the models overpredict current temperature anomalies is not because they are ‘wrong’ but more because the radiative physics is better understood yet the coupling to convection and the effect of clouds between the model levels or grid points (currently clouds are modelled at the grid points) is not. Hopefully a better understanding of the clouds and possible negative feedbacks can be gained and input into the models and we should see better convergence (and possibly a little humility creeping back into climate science).
With regards to the paleo record and also the aquaplanet idea, one of the flaws of this model is that it doesn’t properly represent the inherent mass distribution and gravity field variation that drives thermal transport in the oceans. This is the background that leads to a characteristic frequency signature for the Earth ie ocean tides movements, circulation. A simple example of this is the difference in NH and SH land mass and the behaviour of the Southern Ocean. It is only recently with the launch of missions like GRACE and now GOCE that an estimation of the geoid, and hence gravity gradient, can be determined so that these can be input into the models.
So I would be a bit cautious in trying to relate ice ages to CO2 or negative feedback in the atmosphere before more is known about the underlying frequency space for heat transport. The Earth may naturally go in to an ice age all on its own without any significant external forcing or feedback and this may be in a very short time period. I have spent a lot of my career to now studying material science, maths and plasmas and I see coupled processes and strange transitions all the time; they are actually the most common. To understand coupling you need to understand the components of the coupling reaction first.
Andrew Dessler (#51)
“I guess I have not thought about why variations of the deep time climate are not more variable. One reason is that it seems quite difficult for the planet to get out of a snowball Earth configuration. I don’t think anyone has a good theory for that.”
Difficult? In this configuration, isn’t there a lack of clouds, virtually no open ocean left, and lack of new precipitation? Also, Earth’s orbit is going to likely be in a somewhat different state with the Milankovich cycles. Currently, our aphelion and perihelion incoming power variation, peaking in January, provides peak insolation that is around 95W/m^2 greater at the TOA than the July minimum and we are not currently near a cycle peak as I recall.
With little precipitation, there is sublimation in process so there is going to be some glacier mass lost to that. Volcanic activity is going to put darker ash and dust on the glacier reducing its albedo over time as will whatever meteorite additions. As snow packs down with age, it tends to lose some albedo. As seasons change, cycles change, there will be some surface melting, probably accompanied by refreezing. Then you’ve got potentially clear ice and your albedo has plummeted to around that of liquid water for those spots in areas when there is low angles of incidence w.r.t. the normal (vertical). Now, you’ve got a lower albedo independent of temperature, if your cloud system is still sparse, and it will permit additional incoming power to be absorbed. There are also the occaisional transients that periodically occur which should also help out, the odd comet fragment, asteroid, large meteorite, volcanic erruption will likely play a part over time.
Remember, even an increase in co2 to 4000ppm is going to amount to an additional direct power absorption of under 16 w/m^2 average. The annual orbit variation nowadays is 95W/m^2 TOA which for the short term when averaged a bit means a variation close to that magnitude and the cloud fraction variation during 1998 apparently gave us a longer variation time with an additional 10 W/m^2. Also, the orbit variation is putting that power into the SH, dominated by very low albedo, oceans being close to 0.04 for low incidence versus land that runs many times more, for an average surface albedo of around 0.08.
WRT ice ages and snowball Earth, there are even clouds on Mars. There exist both CO2 clouds and
H2O clouds.
This makes me wonder if clouds would be completely out of the picture even in an ice age or snowball Earth scenario. I think there is just too much we don’t know. Climate models are at best just a guess about how things work. For one thing, life is a huge wild card. It has been a major force in forming our planet to include the climate. It has been resposible for major changes in the composition of the atmosphere. Even in a snowball Earth scenario, some sort of algae or bacteria could significantly alter the albedo.
snip -please do not editorialize about climate models. Let’s stick to cloud parameterizations.
As far as I know, most GCM’s use one sensitivity (within small borders) for all types of forcing, regardless of the completely different actions that the different forcings show. For solar, the main actions are in the stratosphere (UV – ozone) and the oceans (deep penetration of the surface layer) mainly in the tropics. GHGs have their main action in the (lower) atmosphere, more spread over the latitudes and IR doesn’t penetrate the ocean’s surface layer more than a fraction of a millimeter (higher temperature of the skin/evaporation/higher radiation level). Some differences in action of the different forcings on cloud cover may happen directly (absorption/reflection) or indirectly (jet stream position shifts during a solar cycle – long term shifts?). Thus, why can’t there be a strong negative cloud feedback for greenhouse gases (as lower tropospheric process) and a much smaller negative (or even positive) feedback for solar? That should give a higher sensitivity for solar and a lower sensitivity for greenhouse gases…
Thus instead of
dT = S*(dF1 + dF2 + dF3 + dF4 +…)
one will have:
dT = S1*dF1 + S2*dF2 + S3*dF3 + S4*dF4 +…
It works anyway for the past century: one can halve the sensitivity for CO2 and reduce the sensitivity for human aerosols to 1/4th, which explains the past temperature as good as the original “one sensitivity” calculation. See:
http://www.ferdinand-engelbeen.be/klimaat/oxford.html
The necessity for a huge extra help from greenhouse gases to explain the ice age / interglacial transitions then doesn’t apply anymore: with a higher sensitivity for insolation changes and ice albedo changes, due to cloud feedbacks, there is no need for a high sensitivity for greenhouse gases…
snip – CO2 lead-lags have been discussed endlessly. Save it for another occasion.
snip – we’re trying to get information on the conventional viewpoint right now. Let’s stick to elucidating that.
snip – personal viewpoint
snip – personal opinion
Folks, I’m sorry to be snipping but it’s important to try to fully understand what the conventional viewpoint is before getting all excited about alternative theories. For present purposes, look at Hoffert and Covey; Annan and Hargreaves or that sort of thing, not Spencer. If you can see some issues with the texts relied upon by IPCC, fine. But please examine these articles and for discussions here, try to stay away from “skeptic” literature.
Re: Steve McIntyre (#88),
“try to stay away from “skeptic” literature.”
Steve if we were all to stay away from ‘skeptic literature’ as you suggest then we wouldn’t be visiting this blog very often would we? I know you’ll disagree with labeling CA as ‘skeptic literature’ but the fact is, that is exactly how CA is perceived by warmers i.e. as the skeptics web site.
While I appreciate that you want people to stick to discussing the subject of the thread, I think you are being very harsh these days in snipping a lot of peoples post for ‘editorialising’. Please appreciate Steve that CA is an important resource for presenting ‘alternative’ view points which just as easily explain recent millenial timescale climate changes as does the GCM embodied CO2 theory of the warmers. While I agree with you ‘that is important to understand the conventional viewpoint’, its also equally important consider alternatives as well particularly if they tie in better with the observable evidence.
KevinUK
The Egyptians in pharonic times knew that clouds warm at night. To make ice, they would dig a well, insulate it with straw, put a pan of water in the bottom and collect ice in the morning.
If the night had been clear.
Deep space is effectively a black body, and if nothing (like a cloud) interferes, heat will follow the Zeroth Law and dissipate.
Steve, I was not aware of existing of “skeptic” literature as a technical term. Please, inform Dr Spencer that since he is a “skeptic”, his papers are forbidden to be referenced as evidence for everything. You snipped my previous comment asking me to provide citations as support, and then comment on them, not just to express my personal opinions. After doing exactly that, providing reference to Spencer’s published work as support for my previosly snipped claims, I have learned that not all citations were allowed, but only “non-skeptical” ones. So, I would be glad if you send me your Index Librorum Prohibitorum (non-Skepticorum:)) to know what references are allowed, i.e. “non-skeptical”.
Steve: Editorially, I’m trying to focus in this thread on mainstream literature which derives the stated sensitivities. Until those arguments are understood, there’s no point examining other theories. It’s the same thing as I do in proxies. I look at mainstream literature – Mann, Briffa, etc. It doesn’t mean that I always agree with them. 🙂
I’d bet there is no “mainstream literature which derives the stated sensitivities.” There is no analytical derivation. The well-understood radiation physics is not a climate model, and so is insufficient to any conclusion. Sensitivities are found by using GCMs to model climate events, with the conventional forcings adjusted until the event is wiggle-matched by model output. Climate sensitivity is then inferred from whatever was required in terms of parameter adjustments. The confidence placed in this sort of pseudo-experiment follows from Jerry North’s claim that, “we know all the forcings.” If it is assumed that the models are complete, then the outcomes are accepted at face value.
.
However, we don’t know all the forcings. The semi-empirical approach to climate sensitivity therefore rests on circular thinking: Assume the models are complete (the GN error) — adjust the models to reproduce an observable — reproduce the observable — successful reproduction validates the model — therefore the model is proven complete. And so climate sensitivity is now known.
.
All of this plays right into Jerry Browning’s objections about non-physical models with necessarily non-physical parameterizations. Most climate scientists seem to be judging the models by whether their mean outputs correspond to climate observables. Some of them have posted here. But the proper way to judge the models is by the quality of their internal physics. None of the climate scientists who have posted here, expressing confidence about sensitivities, can give an analytical error bar on their judgments. They express confidence without limits. Sorry to say, this is not science.
Steve:
“But please examine these articles and for discussions here, try to stay away from “skeptic” literature.”
Are these papers readily available? Peter Tillman asked for a link – I haven’t seen it yet. Now in my case, I probably wouldn’t be able to understand most of them, but I would be interested in seeing what the other, far more knowledgeable people here have to say.
You recommend Ou and quote him as saying “rather than a low cloud that cools the surface, it is the surface temperature- somewhat independently constrained by water properties- that determines the cloud covers through global heat balance.”
I may be well off base, but isn’t this Dr. Spencer’s point?
Re: Hemst 101 (#92), As far as a link to some of the papers, the best I can do is get the H&C Abstract from Nature:
http://www.nature.com/nature/journal/v360/n6404/abs/360573a0.html
There is also a similar paper by some of the same authors 4 years later which appears much more sophisticated is available from SpringerLink:
Click to access K46605625308276X.pdf
If anyone can access these papers and given those of us without subscriptions and what not any insight into their details, it would be greatly appreciated.
Steve,
I am afraid we have misunderstanding. I was not using Spencer references to propagate his general views on global warming or even to discuss his theory per se, but just to criticize the other theory by using his insight about natural cloud variability for which I think is correct (whoever formulated it).
Thanks for guidelines concerning ILP 🙂
I know that readers are interested in Spencer and so on. And that’s fine. It’s not that I’m uninterested in Spencer’s view. It’s just we all know where to find it.
However, if readers want to understand the context of the debate, then they have to thoroughly understand the arguments of the other side. Are there assumptions in Hoffert and Covey that may not hold up? Or are their arguments convincing? Same with Annan and Hargreaves.
I think that it’s important to dissect the literature – I’ve obviously done so in other areas. If people want to dislodge these points of view, then they have to understand them – it’s not enough to merely assert an opposite theory.
And unless people actually quote from the literature being criticized, they generally haven’t come to grips with the exact argument.
We have this same problem with any physics related thread, which is one reason why I seldom encourage such discussion.
I don’t have Hoffert and Covey but I’ve placed Covey et al 1996 online here
The opposing position to Covey et al 1996 is Lindzen. A comment on Hoffert and Covey in 1993 here . Lindzen and Pan 1994 http://eaps.mit.edu/faculty/lindzen/171nocephf.pdf . Covey et al 1996 also mention a 1993 AGU talk by Kirk-Davidoff and Lindzen – I can’t locate whether it was published or not,
Covey et al 1996 discuss several Hansen papers:
Click to access 1993_Hansen_etal_1.pdf
http://pubs.giss.nasa.gov/docs/1985/1985_Hansen_etal.pdf climate response times
http://pubs.giss.nasa.gov/docs/1997/1997_Hansen_etal_2.pdf wonderland
Knutti and Hegerl 2008 is online here.
Annan and HArgreaves http://www.jamstec.go.jp/frcgc/research/d3/jules/GRL_sensitivity.pdf
Another article in the same vein: http://researchpages.net/media/resources/2008/01/11/Edwards_et_al_2007_PPG_.pdf
Kenneth Fritsch:81
It might be timely to remember that there are 2 UN bodies under the UNEP structure.The IPCC and the UNEP ozone assessment committee.
Here interpretations of the same papers have different perspectives.eg Larsen 2005 cited under both “umbrelleas”
eg Larsen 2007.
UNEP. (2008). Environmental effects of ozone depletion and its interactions with climate change: Progress report, 2007. Photochemical & Photobiological Sciences 7: 15-27, DOI: 10.1039/b717166h
Re: DeWitt Payne (#78),
tnx for the reference.
I didn’t quite see that a brief reviewing of the article. The change in nucleation particle size assumed reduced the effectiveness with less cloud albedo. Albedo was reduced from today’s 0.30 down to around 0.22 with a reduction in cover from 64% down to 55%? or with a 27% reduction. Eyeballing the charts looked like the T differed by about 5 degrees for 4xc02 and about 15 deg C rise for both 4xc02 and the change in cloud cover due to the nucleation particle size increase.
The increase in power absorption, clear sky, for co2 4x is around 7.4w/m^2 while the increase in insolation due to a drop of 0.06 in albedo is around 20.4 w/m^2 or roughly about 28 w/m^2. A sensitivity of 15k/28w/m^2 is about 0.5 K / W/m^2. That’s a little less than the 0.7k/w/m^2 of the 4xco2 only given.
Judging by the graph maps, it looks like there’s a good bit of difference in continental position. I’m not sure how much the disparity is between NH and SH ocean vs land here but it too looked like it might be a little more even than today.
As for the positive/negative feedback nature, that’s both a matter of timing (day/night presence), and the thickness and nature of the cloud type and what the average cloud effect turns out to be. The paper is achieving the massive temperature increase by the 64 to 55% reduction in cloud cover. However, I wasn’t able to get a number other than that of the delta albedo which amounted to 27% of the total cloud cover albedo. It would seem that changes in the timing for the clouds would also contribute more positive feedback if they extend the cloud life into the night. That would be some unrecognized additional forcing (on my envelope back anyway) and there would be a lower sensitivity. Less nighttime cloud cover would result in higher sensitivity than my above one.
I’m confused about this thread, Steve. The work you originally cited regarding Bretherton’s climate sensitivity work — and more recent work by them I blogged on 6 months ago — would seem to support a “skeptics” argument. So why are you asking people to understand the mainstream of climate sensitivity work when that is what we get inundated with, ad nauseum?
I predict that, until people start distinguishing cause from effect in cloud behavior, there will be little progress in cloud feedbacks. The causation issue was the first thing that occurred to me, as a meteorologist, years ago when I asked why researchers thought cloud feedbacks were positive. Now it has been published by us, Nov. 1 2008 J. Climate, and the two mainstream IPCC people who reviewed the work agreed we raised a legitimate issue. One said it was important for other modelers to be made aware of the issue.
For some reason, I have had ZERO input since (unless you count being called a “denier”, in which case I get input routinely). 🙂
Re: Roy Spencer (#100),
Hi, Roy.
I think that it’s important for people to understand the context of the debate – especially if you come at it with a bias. Obviously there are a lot of skeptic readers here and that’s fine. But everyone has to be aware of their bias and be careful of articles that they “like”. If I see some proxy series with a high MWP, I have to be careful about the same sort of bias that I criticize the Team for.
I think that the cloud sensitivity issue is very important and obviously you’re at the front edge of this debate.
But it’s important for readers here to understand where Dessler, for example, is coming from. For exactly the same reason as I want to understand the mechanics of Mann or Steig or whatever. Right now, it seems to me that the lead bloggers have a better understanding of Steig than the authors.
Dessler’s argument against low climate sensitivity is paleo. OK, I happen to like paleo. This leads back to the articles cited above and to the Lindzen debate. I’d like to get these into context before coming forward to your articles – so readers – and myself – who aren;t intimately familiar with the issues as you are can place them in context.
It’s not that we won’t get to the current work, but Dessler’s defence of high sensitivity seems to be based on the older Hansen/Covey work, so I, for one, want to be sure of the footing.
It makes sense to me that cloud forcing is somewhat negative. But Dessler’s got a point about the ice ages as well. But it’s just a point. I’ve perused a few of the cited articles and the issues vis a vis Lindzen don’t seem to have been resolved. (They are a different Lindzen issue than iris). It looks a bit as though they just asserted things forcefully over and over as though that were proof – the same sort of thing that we see with the Team on proxies. But the issue from the mid-1990s needs to be dissected a little.
I don’t know why everyone is so impatient with this approach. It gave a lot more insight into the MBH hockey stick than people merely presenting squiggles with a MWP, which is what people prior to Ross and I did.
Re: Steve McIntyre (#101),
Have you offered Dessler a chance to make a guest post here precisely to provide what you want; an overview of what the paleo data suggests and how it’s either entered into or derived from the climate models? I’m a bit worried about expectation bias entering into the modeling, but perhaps there are ways being used to protect from such bias.
Re: Steve McIntyre (#96),
Consider
Re: Andrew Dessler (#30),
I’ve quoted it and it seems to a basic argument of Andrew’s rejection of the Cloud Super-Parameterization, but there’s no physics or even any logical argument here, merely an argument from inability.
The paleo argument we’ve been through here a while back. I didn’t buy it then and don’t buy it now. As Roy Spencer says above, it’s a matter of causation. I don’t see any way to prove that higher levels of CO2 are the cause of higher temperatures as opposed to the result of it, when we’re talking millions of years ago.
Re: Dave Dardinger (#102),
DAve, you rightly raise your eyebrows at Dessler’s reasoning:
A mathematician wouldn’t use the following argument:
For third parties wishing to understand this argument from first principles, the elephant in the room is, of course, the lack of any physical theory (or more precisely, the lack of any generally accepted physical theory) explaining why CO2 changes so dramatically over Milankowitch periods. Amplification of changes by CO2 is hardly an implausible idea, but without an accepted and plausible explanation for why CO2 varies so much, it’s a very incomplete argument. Having said that, being consistent, just because people haven’t explained it so far doesn’t exclude the possibility that someone will come along and provide a good explanation.
The other defect in Dessler’s argument is that in the early 1990s, Lindzen wrote some papers proposing a precession-related mechanism for amplifying Milankowitch factors. These may or may not be valid, but it’s not impossible that Lindzen could have figured out something that North couldn’t.
Knutti and Hegerl 2008, the review article on sensitivity referred to us above, has a section on the paleoclimate argument. It cites Hoffert and Covey 1992 and Covey et al 1996, but not the opposing Lindzen articles – the sort of the thing that one unfortunately has to watch out for far too much in this field. I’m not familiar enough with the subsequent literature to know what the counter-arguments are. Sometimes, as CA readers know, the debate may have been nothing more than people repeating the same thing over and over again in ever louder voices – a practice that is not “unprecedented” in climate science e.g. Jolliffe’s review of the MBH reply for Nature in 2004. If so, then it would be worth re-examining the Hoffert-Covey v Lindzen issues in light of the subsequent 14 years – the sort of consideration that one would have expected from Knutti and Hegerl 2008 in Nature. However, as we’ve observed before, the quality of reviewing of climate science articles in Nature is not very demanding.
If, as Lindzen argues, there are other important amplifying factors covarying over a Milankowitch period, then this has a material impact on the arguments in Hoffert and Covey 1992, Covey et al 1996 or the Hansen papers. Each of these papers uses the simplest possible comparisons – I’ll post on this sometime. While I can wade through esoteric multivariate methods, I happen to think that it is usually a good idea not to shy away from using simple statistics as they can keep people on track. However, if Lindzen has succeeded in identifying a covarying method of amplification, then the very simple statistical analysis of the above papers probably doesn’t hold up anymore without a demonstration that the Lindzen covarying factor is quantitatively negligible.
Re: Steve McIntyre (#103),
There is a strong (surprisingly linear) correlation between temperature and CO2 levels of about 8 ppmv/°C over the whole 420,000 years Vostok trend, with a lag of a few hundred to a several thousands of years for CO2 after temperature increases. This largely (but not completely) follows the temperature/solubility curve of CO2 in seawater. A complete transition of an ice age to an interglacial needs about 5,000 years, the opposite about 10,000 years. Thus in most cases the lag is small enough to allow for a feedback of CO2 (and other greenhouse gases), increasing the initial small temperature increase from the Milankovitch cycles. That is the basic background theory of Hansen and others to explain the total increase/decrease in temperature. According to the models, the ice/land/ocena/forest albedo changes are not strong enough to explain the total warming/cooling.
See: http://pubs.giss.nasa.gov/docs/2003/2003_Hansen.pdf (page 6 gives the estimated forcings). CO2 is good for about 30% of the change, the rest of the greenhouse gases about 10%.
Because of the overlap, it is difficult to prove/disprove the CO2 feedback. Any alternative amplifying factor (like clouds), or an underestimate of the sensitivity for solar/ice albedo changes would reduce the need for a CO2 feedback.
A detailed observation of the temperature – CO2 increase between the LGM and the Holocene, based on the Dome C ice core, doesn’t show any feedback of CO2 on temperature, if one looks at it through engineering glasses, while a 30% positive feedback should be highly visible. See http://gallery.myff.org/gallery/145233/epica5.GIF with thanks to André van den Berg who supplied the graph.
Re: Ferdinand Engelbeen (#105),
Ferdinand, thanks for the link.
Readers, PLEASE do not use this as an occasion to debate CO2-temperature lead-lags. It’s an important issue but it’s been discussed endlessly and editorially I don’t wish to go there in this thread. There are other issues here that are less explored/
Re: Ferdinand Engelbeen (#105),
Please see my reply to your last paragraph in Unthreaded – n+2.
Re: Steve McIntyre (#88)
Folks, I’m sorry to be snipping but it’s important to try to fully understand what the conventional viewpoint is before getting all excited about alternative theories. For present purposes, look at Hoffert and Covey; Annan and Hargreaves or that sort of thing, not Spencer. If you can see some issues with the texts relied upon by IPCC, fine. But please examine these articles and for discussions here, try to stay away from “skeptic” literature.
Steve. I suspect you wrote this one in haste. In your other statements about the importance of the debate, you add,
“I think” or “To understand” and other qualifications. In this case you state boldly “it’s” and I think this is short sighted. It might or might not be important to fully understand what someone else says, but in the quest for the truth I suspect and would like to submit for your consideration, this is not always the case. If someone else finds the truth and
presents it, they have no obligation to read, discuss, show why the other side it wrong. If they do all the better, but just finding and stating the truth is worthy enough. By the way, I work for a major car company in the environmental dept. and my company will spend a great deal of money on global warming issues whether there is a shred of truth to the anthropocentric aspect or not. I frequently distribute your presentations to our staff, and I am thankful that you
have done the heavy lifting on so many topics. I apologize that my lone submission is dang near trivial.
Lindzen treated the climate models as a black box, which from what I gather to varying degrees they are, and derived the sensitivity from the predicted incoming and outgoing radiation. This strikes me as a clever method to circumvent the details of the model and still derive a critical parameter/factor/whatever for comparison to the measured radiation input/output. Do you who study these sorts of things every day agree with Lindzen’s conclusion that the observed sensitivity is negative? snip
Steve: most climate scientists disagree with Lindzen. I, for one, don’t study this particular “every day”. Indeed, it’s mostly new to me as I’ve said here.
Re: Jim (#107), Silver bullets don’t work on vampires, only werewolves. Also, Lindzen’s approach is interesting (ERBE is well worth analyzing for sensitivity information, especially if we make use of phase-space analysis, as Spencer is doing) but totally unrelated to the present discussion, which is the paleoclimate argument for high sensitivity. Lindzen’s argument is that the simplified approach to test sensitivity doesn’t work with non homogeneous forcings (like Milankovitch), the counter argument seems to be that there is no way to test the hypothesis that the spatial heterogeneity of forcings matters for paleoclimate. This all related only tangentially to his feedback work.
Re: JamesG (#109), They present opinions, yes, but we want to understand there opinions and analyze their validity without dismissing them out of hand. So let’s focus of the paleo papers that argue “mainstream” results for the moment. Who knows? They may be so flawed that rebuttal isn’t necessary. Certainly CA has found some messy paleo work in the past. Heck, that’s why CA exists.
Re: Andrew (#110),
I didn’t dismiss the opinions, just the Bayesian approach which in A&H leads to a pseudo-Gaussian distribution centered on 3 degrees C. This idea has come to be the standard accepted view of the sensitivity but it is flawed if Bayes is inapplicable, which is the case. It also becomes a self-fulfilling prophecy for climate model tuning. There are far more conventional approaches to this type of uncertainty but as they are not relied upon by the IPCC I won’t mention them.
Re: JamesG (#143), Apologies, I think you misunderstood what I tried and failed to convey, which is that Steve wants us to focus on the standard paleo “mainstream” argument and I think that we ought to do so as well. It should be illuminating for everyone. No slight intended against you at all.
Nature does not parameterize, modellers do. Appreciation of that profound distinction is little in evidence here. If my repeated plea (#79 & #86) that we should study nature itself more thoroughly when key physical knowledge is sorely lacking and confusion between cause and effect is rife strikes you as no more than personal opinion, then so be it. Not everyone can be expected to be versed in the physics of nonlinear systems. I hope, however, that you do not intend to turn Climate Audit into the Review of Climate Team Literature. Many seasoned professionals who analyze and model field measurements have already been there. They will only yawn.
P. S. As I write this, the coastal zone of Southern California is socked in by a seasonally typical layer of clouds, the “June Gloom” that depresses temperatures several degrees C from clear-sky conditions.
Finding logical flaws in the existing theory is not the same as presenting “fringe” theories and Lowell Stott and Jeff Severinghaus are certainly not skeptics or alternative theorists. They are indeed utterly mainstream which is why I quoted them.
But you snip opinion when opinion is all there is. As Annan and Hargreaves state quite well in their paper there is no real calculation of sensitivity without masses of subjectivity. All A&H did was collate this subjectivity in a Bayesian framework. But Bayes only works for expert opinions which have been proven to be reliable, so that approach is worthless. Even Dessler forced the point home that it was a subjective issue. You’re seeking something that just isn’t there. How can anyone clinically dissect the ethereal? Only skeptics like Lindzen, Christy or Spencer have actually compiled a sensitivity estimate using real data.
Nobody has any good explanation for the ice-age cycles. That much is well understood by everyone. But the point about CO2 being a heating amplifier is that everyone stops there and completely fails to discuss exactly how it magically switched to becoming a cooling amplifier. The reason is simple – because there isn’t any mechanism for the switch. I defy anyone to find any paper that describes such a mechanism. Well ok, there is one which talks about a seismic switch but it wasn’t a good argument which is why it’s ignored. The truth is that if you are a true believer in AGW then you don’t really need an explanation for everything – about 50% is good enough.
Covey et al 96 was interesting but troubling.
Their table 1 of Q’s and T’s was similar. While it’s all mostly guess work, it looked like the dQ/dT error analysis was also guesswork. It looked very sloppy, which is hard to fathom at such a low resolution.
for a 2x co2, they claim 4 w/m^2 additional forcing from today’s values. I seem to recall seeing 3.7 W/m^2 being the most common value being bandied around and both the online modtran calculator and my own 1-d model offer up around 3.5 w/m^2 for the 10-11km tropopause value for an average (1976 standard atmosphere).
Figure 1 dT/Dq 2x co2
H&C LGM 2 +/- 0.5
H 93 LGM 3 +/- 1
H&C Cret. 2.5 +/- 1.2
Barron 93 3.8 +/- 2
Just doing the calculations min/max/midpoint/range resulted in:
H&C LGM 1.9 +/- 0.6
H 93 LGM 3.1 +/- 1.2
H&C Cret. 2.5 +/- 1.8
Barron 93 4.9 +/- 3.1
While doing the same using 3.5 W/m^2 tropopause value for a doubling (384-768ppm)
H&C LGM 1.6 +/- 0.5
H 93 LGM 2.7 +/- 1.1
H&C Cret. 2.7 +/- 1.6
Barron 93 4.3 +/- 2.7
H&C evidently used a surface albedo variation in addition to their co2 variation for the Cretaceous. It was 6W/m^2 worth of increased power due to a decrease in surface albedo. Crudely, it was around the equivalent of dropping all modern land glaciation from modern values. This resulted in a slightly less co2 sensitivity.
NOWHERE that I read was there any attempt to analyze or quantify anything other than CO2 and surface albedo. There was no factor thrown in for other ghg’s such as methane or variation in cloud albedo. It was assumed there were virtually no deserts so mostly there was supposed to be tropical forest cover, I’m sure complete with serious amounts of methane producing insects and decaying vegetation that raised the methane levels substantially over today’s. Considering there are concerns over positive feedback with clathrates and problems being described with bovine flatulence, failing to account for additional methane seems to be a gigantic flaw in the measurements.
The addition of more rain forest suggests potential serious changes in cloud formation and the presence of much more h2o in the atmosphere that is not necessarily just some feedback of temperature, and hence co2. There may be a positive water vapor feedback from the additional temperature and presuming a constant relative humidity, but biological associated water vapor, jungle versus desert, not solely a function of temperature.
It doesn’t look like this additional water vapor is really amenable to detailed speculation. The vapor itself is high absorption in the IR. How clouds form from it and when they’re present, day and or night, is going dictate the positive/negative forcing nature of it and ultimately, whether it will be positive or negative. Combine that with just how much of this added water vapor is a feedback from temperature versus present because of the land usage, jungle versus desert is putting things more into the realm of idle speculation.
For the sensitivity, the only factors that may be accounted for are the CO2 levels, surface albedo, and possible water vapor feedback which even excludes added water vapor due to land usage. All other factors, such as additional methane as well as the non feedback water vapor are ignored, leading to increased estimates of sensitivity. The factors of atmospheric albedo variations, night time clouds generating positive forcings are totally left out of the analysis.
It’s no wonder why the sensitivity is so high.
Re: cba (#112),
To be fair about this sort of study, there’s only so much information available available. This sort of exercise is inevitably going to be guesswork to a considerable extent. As long as the people say so – and I think that Covey et al declare their assumptions quite reasonably – I don’t think that it’s reasonable to say that it’s “sloppy”. It’s an attempt to do a broad brush scoping. It is what it is and I don’t criticize the authors on that count.
This sort of observation seems more like a debating point to me. No one’s established that bovine flatulence is a material issue in the present debate and I doubt that it is. So I for one am not going to rag Covey et al for not considering that sort of issue.
As always, I encourage people to read the relevant literature before venturing their own bright ideas on things. Lindzen wrote articles that are exactly on point and offered up his own theory for a mechanism for effecting ice age changes with low CO2 sensitivity. I urge readers to familiarize themselves with Lindzen’s ideas before venturing their own “bright ideas”.
Why do persistent errors arise in Paleoclimate studies and indeed test from affixing arbitrary stationary temporal points in orbital forcing ? Such as the albedo limit at 65n in the “fixed summer solsice” eg Crowley and North, 1991
This is clearly not the case eg
Temporal drift
Budyko, M.I., 1977. Izmeneniye klimata Climatic change.. Gidrometeoizdat, Leningrad, 280 pp.
Sloan and Morrill described “persistent discrepancies” between climate model results and interpretations from proxy data in the Eocene.
‘Essential for development and testing of new models of Pangean climate requires proxy data with appropriate spatial and temporal scales. Sloan and Morrill (1998) pointed out the continuing discrepancy between global climate models (with higher pCO2) and geological and paleontological climate proxy data from times of “extreme climate”, such as the Late Triassic and Early Jurassic. They show that orbital forcing of climate can play a critical role in continental climate with extreme values of orbital parameters reducing the interior annual temperature range by 75%, resulting in cooler summers and warmer winters. As shown by Sloan and Morrill (1998), these orbital variations must be taken into account in comparing paleoclimate models to climate proxy records. While the latter comparison requires the specification of one orbital state, the geological proxies span many orbital cycles as well as the full range of orbital forcing. Indeed, it is quite reasonable to expect model-proxy comparisons to be valid only over intervals of time representing one orbital state or minimally the climate proxies should be drawn from homologous portions of several cycles (e.g., times of high insolation). Hence, paleoclimate proxy data from the geological record must be placed in a temporal framework appropriate for Milankovitch-scale modeling. While much remains to be done with outcrops, the workshop panels concluded that very long sections, such as those available through coring provided the best means of obtaining the needed high-resolution data for the next generation of models and model-data comparisons.
Sloan, L.C., Morrill, C., 1998, Orbital forcing and Eocene continental temperatures: Palaeogeography, Palaeoclimatology, Palaeoecology, v144, p. 21–35
Steve (114)
Keeping it down to only one or two significant figures was not my complaint. The data is really rough. It’s being off in the math for the calculated value at the first or second significant figure when providing the calculated value and accuracy range that irked me a bit. Also, providing essentially a known or accepted value to 1 significant figure that is rounded up and then used with two significant figure values was also a little bit over the top to me. The “sloppy” I reserved for their calculation based upon their supplied numbers and errors and their choice of significant figures. Specifically, the H&C Cretaceous error provided was +/- 1.2 yet the min/max error I calculated from their’s showed +/- 1.8 and there was no appeal there to independent error calculation in quadrature as was done later for something else.
My reference to bovines is strictly referencing modern arguments being made concerning GHGs other than CO2. It’s purpose is to suggest that biological methane from whatever source is under consideration as part of the current GHG problem and has been omitted (ignored) in the Cretaceous period sensitivity of the paper despite the possibility that there was a much larger atmospheric fraction then from all sources, natural and biological. I also don’t recall seeing any means of ascertaining what that concentration fraction might have been. What’s important in my view is that the paper doesn’t address the possiblity that the methane levels were higher or the possiblity that water vapor concentrations were higher because of the presence of more jungle and less desert rather than increased water vapor solely being a feedback from temperature. These are factors then that are missing from the dQ, but they are not missing from the dT, creating an artificially high sensitivity. As I thought I understood from the presence of those clathrate methane deposits, they formed, coming out of the atmosphere, and that they are quite unstable and will potentially go back into the atmosphere as the temperature rises further. In either case, we have more forcing, W/m^2, going on with these paleo estimations than the allowable estimated CO2 range along with some possible bit of surface albedo change in some of the studies that must go into the dQ or the result will be a significant over estimate of the sensitivity.
It may not be possible to even estimate these factors. If they are not a factor, which seems incomprehensible to me, then they won’t affect the outcome and are inconsequential. If they are a a factor and have not been considered, then the results of the Covey paper grossly overestimate the real sensitivity.
Sorry for misconveying some notion that bovines were important in the Cretaceous or important now for that matter.
Steve
I think my argument is slightly different than you described it above with your Fermat example. It is not just that Jerry could not explain the Ice Age with present-day carbon dioxide, but that when he put pre-industrial carbon dioxide into the model, the model was able to grow ice sheets. That’s a stronger argument in favor of the mainstream estimate of climate sensitivity.
When thinking about climate change, my views are based as much as possible on the big picture. If you just look at a single data point, you can always find some alternative explanation that cannot be excluded, or something about it that doesn’t quite fit, etc., but I think that gives an incorrect view. You have to look at all the data — the present-day warming, ice ages, the Eocene, etc.
If I may utilize a statistical metaphor, imagine trying to fit a line through data. If you only look at a few data points, you can fit all sorts of lines through the data. However, by adding more data points, the correct slope of the line emerges. Thus, I think that my views are not just based on the paleo data, but based on everything that I’ve seen — modern data and paleo and everything in between + my intuition.
As far as Roy Spencer goes, I give him a lot of credit and think his ideas are very interesting. However, I think it’s also important to realize the limitations of his analyses. For example, his paper on the negative cloud feedback only looked at the latitude range from 20°N to 20°S. It says nothing about what’s happening outside that latitude range. Thus, I don’t think it’s fair to conclude that he has found a “negative cloud feedback.” That shouldn’t be taken as a criticism, since most science is incremental. And I view his work as an increment on the way to an observational estimate of the cloud feedback.
Thanks!
Re: Andrew Dessler (#120),
“I think my argument is slightly different than you described it above with your Fermat example. It is not just that Jerry could not explain the Ice Age with present-day carbon dioxide, but that when he put pre-industrial carbon dioxide into the model, the model was able to grow ice sheets. That’s a stronger argument in favor of the mainstream estimate of climate sensitivity.”
Could he also grow ice-sheets under the conditions prevailing during the Oli-1 sand Mio-1 glaciations, when CO2 was appreciably higher than today?
Re: tty (#122)
When it comes to CO2 data from that far back I believe you can pick a proxy that suits your pet theory:
Linkages between CO2, climate, and evolution in deep time Dana L. Royer
http://www.pnas.org/content/105/2/407.full.pdf+html?rss=1
Covey et al 1996 state:
The citation is to the 1993 AGU conference:
Kirk-Davidoff’s CV reports:
Lindzen’s list of publications says:
You’d think that they’d have fished or cut bait on this article by now, but it’s still “in preparation”. Maybe they’re waiting for some data from Lonnie Thompson.
It seems to me that a model with negative feedback in both the heating and cooling processes could replicate a system that shifts between relatively stable warm and cool modes.
Condider the model for the rate of change of global heat content H (a minor extension of lucia’s model lumpy; ie of the conventional viewpoint!)
dH/dt =(a-b*H)*TSI – (f-g*H)*H
The first term on the right is heating from Total Solar Irradiation, modified by the term a-b*H. This is a “negative feedback” (in the ClimateScience sense). With b>0 an increase in H decreases warming due to TSI (eg due increased H leading to increased cloud cover). The second term on the right refers to heat loss, basically proportional to H but with a coefficient that again incorporates a negative feedback when g is positive (eg again due to increased H leading to increased cloud cover, but with a different dependence than the TSI term). At steady-state dH/dt = 0 and the equation is a quadratic in H which can be written
(H-Hc)(H-Hw)=0
which has two solutions H=Hc and H=Hw; that is, “cool” and “warm” steady-states.
As for time dependence, writing the de as
dH/dt = g(H-Hc)(H-Hw) g>0
For any intermediate H (Hw>H>Hc) dH/dt<0 and H approaches the Hc steady-state. Also if H<Hc0 so Hc is a stable steady-state. The Hw steady-state is not stable but this points to an obvious defect in the original formulation. The terms in brackets in the de can only be approximately linear in H over a limited range. For example, if the dependence on H is through water vapour, these terms would approach constant values at high enough (but not particularly high)H corresponding to a saturated atmosphere. In that case the de can be written
dH/dt = p*TSI – q*H
with p and q independent of H and a stable steady-state at Hw=p*TSI/q.
The model is too simple to warrant serious analysis as a climate model but I think it shows in principle the kind of beaviour that could happen with negative feedback in both directions. If some external disturbance(to the model; eg heating due to deep ocean volcanic activity)to state Hc resulted in an intermediate value of H it could return to either Hw or Hc. A cooling disturbance to state Hw (eg surface volcanic activity producing aerosols) could also return to either state. Any return to Hw would look like positive feedback. Above Hw it will return to Hw, below Hc it will return to Hc, both negative feedback. So either apparent stability or instability depending on the direction and extent of a disturbance.
In Chapter 2 – Feedback Factors and Radiative Forcing, The 2009 NIPCC Climate Change Reconsidered in Ch 2.1 Clouds provides a brief review of nine earlier Cloud articles. This include pro/con Lindzen’s Iris theory.
An interesting review: Evaluating the present-day simulation of clouds, precipitation, and
radiation in climate models, Robert Pincus, Crispian P. Batstone, Robert J. Patrick Hofmann,
Karl E. Taylor, and Peter J. Glecker, JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 113, D14209, doi:10.1029/2007JD009334, 2008
It appears there is still work to be done to satisfactorily evaluate cloud feedback!
In the wake of Tom Vonk’s (#126) fine expose of climate modelling pitfalls from the standpoint of nonlinear physics, I’d like to add some practical points from the standpoint of system analysis, an issue raised by Jim’s (#107) reference to Lindzen’s conclusions and illustrated by Roman’s tiles (#14).
“Black-box” system analysis is very powerful, as long as the system is linear or nearly so. That is the basic premise underlying all notions of “sensitivity.” It is only in such systems that X input consistently produces Y output, albeit on a frequency-dependent basis–a point seldom appreciated in climate science. However, even the simplest nonlinear systems, those with a memory-less response characteristic, produce time-average outputs that do not correspond neatly with the time-average input. The outputs become range-dependent. In cases where the nonlinear characteristic is well-known, so-called “describing function” methods nevertheless can applied to obtain adequate spectral transfer functions for the output.
Our climate system satisfies none of the above conditions. It is highly nonlinear, with abrupt phase changes of water playing a major role in its response. Because the oceans are great capacitors of heat, the system has memory. Many important variables are temperature dependent. The semi-permanent surface albedo varies seasonally. The energetic driver of the whole system, of course, is thermalized insolation. Aside from stratospheric ozone, GHGs do not thermalize solar power. The persistent thermalizer is the ocean-and-land surface of the globe, with episodic minor contributions by clouds. Evaporation and sublimation are very powerful cooling mechanisms (~590 calories are extracted per gram of water) whether they operate at the surface or aloft.
It is the condensation aloft into clouds that is the Joker in the pack. Cloud albedo acts as a flickering gate or grating, sporadically reducing insolation at the surface (pyrometers frequently register drops of 100W/m^2 or more), thereby reducing the thermalization. And GHGs only operate on the thermal power radiated by the surface, producing none of their own. They merely redistribute that power via isotropic radiation and molecular collisons with the “inert” gases that constitute ~98% of the atmosphere, without any amplification. But clouds also are opaque to IR radiation. While they block the radiative path to space for outgoing radiation, they also block the path to Earth for the latent heat released upon condensation. In marine environments the power of the latter exceeds that of outgoing radiation.
Adding to the complexity of system response are the pressure gradients created by uneven thermalization of the surface. They drive the winds, waves, and currents throughout the globe, with a coarsely estimated average of ~30W/M^2 or more converted thereby to mechanical power. There are strong regional and seasonal differences, however. Thus when sunshine hours are regressed upon temperature in England, which benefits from Gulf Stream heating especially in winter, all the different factors come into play in determining the apparent regional “sensitivity.” The scatter of data points is least in summer, when insolation is at its highest. Correlation falls close to zero in January, however, when leaden skies prevail. Such pitfalls are inherent in considering climate “sensitivity” not only regionally but globally.
Climate models have a long way to go before they can adequately represent the manifold complexities of the system. And the problems are not simply those of coarse mesh-size in a finite-element scheme, which prevent them from representing tropical cyclones and thunderstorms. What matters most in determining the variation of global surface thermalization is tropical albedo. Convective development everywhere has a pronounced diurnal variability. Quite apart from the insurmountable chaotic aspects, the crucial mechanism of cloud formation and dissipation thus is not amenable to time-invariant modeling approaches. What is required is a thorough evolutionary analysis of total power fluxes that determines the potential temperatures and cloud cover aloft.
Nonlinear system dynamics simply cannot be identified or analyzed by examining data over climatic time scales. Fundamental knowledge of the underlying physics is required. Alas, what one sees in the literature are little more than empirical rules of thumb based on global average humidity/cloud conditions and sensible temperatures. Small wonder that the results are usually as variable as Roman’s tiles. There can be no consensus on this issue in any sandbox.
Re: John S. (#131),
John – Dr. Lindzen’s approach appears to have circumvented the unknowns in the climate system as well as the climate models. Isn’t the fact that he used SST as the forcing take into account the effect of SS warming due to heat in the ocean and the effect of radiation, as well as other sources of heat, e.g. air? That is what was compared to the outgoing radiation, both measured and modeled. This in effect measures the current driving force vs. the current output, so the multi-decade lag factors shouldn’t matter. I realize the climate models are incomplete and even if they were complete that they could not predict the climate at time X, but it seems odd that all the models predict a positive feedback factor while the ERBE calculation yields a negative one. And it may be that some climatologists take the seriously (maybe!), because some of them are now “adjusting” the ERBE non-scanner data. It might be necessary to do this, but why are these things never fixed before someone uses the data to prove AGW false?? ERBE S10N_WFOV ERBS Edition3
Data Quality Summary
Re: Jim (#141),
I’ll have time only tomorrow for a discussion.
Re: Jim (#141),
Because the current system response depends not only on the current driving force but also upon the state of all the system components, the basic problem is by no means overcome. Indeed, the determination of system behavior that is provided by system analysis techniques can never be any better than the specification of the dynamic properties of all the components. Yes, a blind “black-box” approach can provide valuable insights when such specification is lacking. That is usually the case only with quasi-linear systems, however.
Now, taking a global power-budget view of the climate system in its present (as opposed to ice-age) state, it is probably quasi-linear enough to make it amenable to empirical “black-boxing.” After all, the recorded temperature variability during the last century was well within +/-1K. With such inherent range limitation, even strong nonlinearities do not deviate that much from linear, and the response characteristics estimated thereby provide a reasonable first-cut. But such range-localized estimates cannot be expected to apply to modelling situations where ice ages are considered. And in no case can one dispense with the need do the mathematical analysis properly.
The disturbing tendency in “climate science” is to throw all these caveats to the wind in the rush to find answers. Even more reckless is the resort to short-cut methods and fanciful premises, such as “feedback.” Thus instead of determining the coherence-weighted spectral transfer function between, say, proxy temperature and CO2, empirical “sensitivities” are obtained by simply comparing the ranges of the two variables over the glacial-interglacial cycle, as seen with the Vostok data. The results obtained thereby are scarcely indicative of any inherent system response characteristic.
There’s much more that can be said, but my free time is limited and this thread is not the best place for it.
Re: John S. (#153), “Thus instead of determining the coherence-weighted spectral transfer function between, say, proxy temperature and CO…”
What would be a practical approach for determining such a spectral transfer function between a given proxy and CO2?
Re: oms (#154),
A straightforward cross-spectrum analysis can be used to determine the optimal spectral transfer between system input and output, in accordance with the Wiener-Hopf Theorem. Unfortunately, most proxy series are badly aliased due to coarse sampling intervals, and the Vostok data with their highly irregular intervals are not even proper series. A simple “Roman tile” approach on concurrent quasi-millenial rates of change in the Vostok data produces only slightly positive correlation for proxy T on C02, with R^2 values on the order of 0.1. There’s simply nothing solid there to base any conclusions upon.
BTW, if Roman M is still following this thread, it would be interesting to see what the R^2 value is for each of his monthly tiles.
Have a restful Sunday everybody!
Re: John S. (#155),
The Dome C core data has much finer resolution for the transition to the Holocene. It’s still probably too coarse and time alignment between atmospheric CO2 and deuterium isotope ratio shift data isn’t all that good either, but…
Re: John S. (#155),
The R-square values for the plot in RomanM (#14), look like:
JAN 0.0765208
FEB 0.001747862
MAR 0.02797844
APR 0.04687352
MAY 0.07859212
JUN 0.266198
JUL 0.5727077
AUG 0.4887295
SEP 0.0793804
OCT 0.005142079
NOV 0.1128963
DEC 0.09012025
The correlation is strongest during the summer months when the ranges of values for sun are the largest, weaker during the winter and lowest in several of the spring and fall months.
Re: RomanM (#157),
Thanks, Roman, for the monthly correlation values. You’ve shown quite convincingly that wintertime temperature variations in cloudy England are virtually unaffected by weak local insolation.
The striking thing about your tiles is how easily the results could be misused and miscast. Someone could simply average the data slope for all months, ignoring the different correlations involved, and argue that clouds increase temperatures on a net annual basis. Taken to the logical extreme, it would imply that zero thermalized insolation produces the highest temperatures. Ah, the wonders of “feedback!”
Re: John S. (#155)
“A straightforward cross-spectrum analysis can be used to determine the optimal spectral transfer between system input and output, in accordance with the Wiener-Hopf Theorem. Unfortunately, most proxy series are badly aliased due to coarse sampling intervals, and the Vostok data with their highly irregular intervals are not even proper series.”
John,
In that case, how do we approach this problem since it is not a straightforward application of cross-spectral analysis (which most readers on this thread probably realize)?
Re: oms (#158),
Given proper time series of adequate length, cross-spectral implementation of W-H is quite straightforward. And it proves as useful in discerning the lack of coherent, causal relationship between two series as in specifying the positive case.
If, as it seems, you’re asking me for some methodological magic to rescue many proxy series from their manifold failures, I’m bound to disappoint you. I’m not a witch doctor. Nor can I spare more time for this side issue. Sorry.
I still think that the base temperature is set by sun and sea and not by the physics of air and clouds. I think that the physics of air and clouds (more broadly the air circulation systems) modulates the oceanic variations in emissivity just as the physics of the ocean circulations modulates solar variations in energy supply to the oceans.
I saw the oceans and then the air warm from all those strong El Ninos from 1975 to 2000 and at the same time the air circulation systems moved poleward. From 2000 the air circulation systems started to move back equatorward and the warming first stopped and we are now cooling with less (and less powerful) El Ninos.
Recent observations about multidecadal ocean cycles (seperate in each ocean) that seem to operate over 30 to 60 years or more show that net ocean emissivity/absorption does change on such timescales and I believe we have seen enough to link those emissivity changes to air circulation changes and to link them to all observed global temperature changes and regional climate shifts without needing to involve CO2.
Once one accepts significant variations in ocean energy emissivity working with longer term solar variability then the logic shifts away from tropical convection to the wider concept which I describe in a number of articles at
I think the fuller truth is as follows:
1) Day to day (diurnal) temperature disequilibrium is dealt with very effectively by the process of tropical convection.
2) Seasonal temperature disequilibrium is dealt wth by the observed latitudinal shifts in the air circulation systems which we see every year.
3) Longer term temperature disequilibria caused by solar variation, changes in ocean emissivity or changes in the composition of the air are dealt with by additional shifts in the air circulation systems beyond normal seasonal variation.
Combining all three pocesses is needed in order to account for the observed apparent stability of Earth’s temperatures for so long.
The point about CO2 or any other GHGs including water vapour is that they alter the characteristics of the air alone. As you have observed, the reaction of the air above water when warming of the air occurs is so strong (due to the energy values of the latent heats of evaporation and condensation combined with accelerated air movement) that warming of the air alone cannot warm the water.
Thus the only available global response to changes that affect the air alone is to accelerate the extra energy to space by shifting the air circulations latitudinally imperceptibly and thereby maintain the sea surface. surface air temperature equilibrium.
Re: Stephen Wilde (#132),
You’ll find no fundamental disagreement from me. Indeed, solar power is what drives the whole system, and the oceans are the principal capacitor of thermalized energy. There can be no doubt that the atmosphere is heated from below on a global scale. My whole narrow OT point about transient clouds is that they are the principal regulator of how much solar power is thermalized by oceans and land surfaces on a day-to-day basis, which in turn determines the variations of temperature and pressure around base values.
As Geiger pointed out decades ago on p.11 of his classic monograph (“The Climate near the Ground”): “night and day, winter and summer are by no means equally matched opponents.” It is the accumulation over time of these ever-present disequilibria that produces the longer-term oscillations that we call ENSO, PDO, AMO etc. But they, along with atmosheric pattern adjustments, are all heat-expunging responses of the complex nonlinear system and not its ultimate drivers, per se.
Whoops. Something went wrong but the sense is still there. A pity there is no edit facility.
Excellent analysis by Roy Spencer of using ice core record for inferring Co2 sensitivity
Ice Ages or 20th Century Warming, It All Comes Down to Causation
Main conclusion:
Re: Ivan (#135), One teensy weensy problem-Vostok is giving Antarctic temperature change, which was probably on the order of 4 times the global temperature change. Many points he makes are fairly good though.
There is a lot about Lindzen’s paleo arguments here, BTW:
Click to access Lindzen2_maj_06.pdf
Shame there is no audio/video, his remarks accompanying some of the slides would be greatly appreciated.
Re: Andrew (#136), but the Vostok and other inland ice cores dD and d18O include a SH oceans component, thus reflect more than only the local Antarctic temperatures. See the discussion by Jouzel e.a. for corrections of the temperature scale at Vostok:
http://www.ipsl.jussieu.fr/GLACIO/hoffmann/Texts/jouzelJGR2003.pdf
That influences the ratio between CO2 and temperature changes somewhat (but not much, about 10%), and the magnitude of the sensitivity for whatever forcing that causes the temperature increase in the first place.
That doesn’t solve the problem of the influence of clouds (and thus overall or specific sensitivity) during the transitions from glacials to interglacials and back and the more or less steady states in these period, except if (Arctic/Antarctic) precipitation levels and (NH/SH) cloud cover show a (very) good correlation.
Re: Ivan (#135),
And here is Spencer directly on cloud cover:
http://www.drroyspencer.com/research-articles/global-warming-as-a-natural-response/
“Global Warming as a Natural Response to Cloud Changes Associated with the Pacific Decadal Oscillation (PDO)”
The key diagram in this article is Figure 1, which compares IPCC model projections of warming to to 5 alternate empirical projections, including a simple linear extension of observed 20th century warming. All 5 alternate projections imply much lower climate sensitivity than the current consensus 3 deg C/2x CO2 figure.
Spencer argues that
(Source: http://www.drroyspencer.com/2009/05/a-layman%E2%80%99s-explanation-of-why-global-warming-predictions-by-climate-models-are-wrong/)
Spencer doesn’t seem to have gotten any traction with his cloud arguments, so far anyway. In another blog post, he asks
(Source: http://www.drroyspencer.com/2009/06/ipcc-admits-they-could-be-wrong-about-humans-causing-global-warming/)
Dr. Spencer, your comments on this topic are welcome.
Peter D. Tillman
Re: Peter D. Tillman (#137),
Once again, I would like readers to focus in this thread on “conventional” literature that is relied upon by IPCC, rather than alternative viewpoints that are not applied by IPCC.
This is particularly important for readers whose inclinations is “skeptic”. It’s easy to read articles that reinforce your biases, but it’s more important that you understand literature that doesn’t.
Re: Steve McIntyre (#138),
Please note that the Spencer articles support the article you quote above:
Cheers — Pete Tillman
Re: Peter D. Tillman (#147),
Forgot to mention that the Bretherton et al. “Cloud Feedbacks Climate Process Team” have a project page with other recent pubs and docs at:
http://www.atmos.washington.edu/~breth/CPT-clouds.html
— I found Bretherton’s Powerpoint slideshow (3-06) interesting. They also have a Cloud feedbacks reading list at:
http://www.atmos.washington.edu/~breth/CPT-public/CPT-reading-list.html
— which is tougher going for amateurs like me.
Happy reading–
Pete Tillman
Re: Steve McIntyre (#138),
“whose inclinations is “skeptic”. It’s easy to read articles that reinforce your biases, but it’s more important that you understand literature that doesn’t.”
Steve, I agree with your point BUT you are seem to be saying that if you declare yourself as skeptical of the current AGW hypothesis proclaimed by the IPCC then you are biased? If you have a contrary opinion on anything does that make you ‘biased’? If so (which I don’t believe) then I’m happy to be a biased Popperian scientist who suffers from biased thinking just because he thinks that any hypothesis must at least be capable of being falsified before it can even be considered a hypothesis let alone advance on (after it has shown to have merit through repeated failed attempts to falsify it) to become a theory.
KevinUK
Some informative documents:
Seasonal variation of cloud radiative forcing derived from ERBE
Click to access Harrison%20et%20al%20JGR%2095%20D11%2018687-18703%201990.pdf
On the observed near cancellation between LW and SW cloud forcing in tropical regions
Click to access i1520-0442-7-4-559.pdf
Some visuals and background
http://icp.giss.nasa.gov/research/data/erbe/
Click to access week13.pdf
http://cimss.ssec.wisc.edu/wxwise/homerbe.html
Re: thefordprefect (#140),
It is unfortunate these satellites only operated for five and eight years beginning in the mid 1980s. It would have been interesting if we could match up data from ERBE and ARGO to see how any radiative imbalance was warming oceans. Perhaps it would have shed some light on the role of clouds.
Steve,
I mentioned Spencer only because he commented on the issue that was here discussed – paleo arguments for CO2 sensitivity. Basic Hansen’s argument is that since Milankovitch initial forcing was to small, then CO2 feedback must be responsible for most of the interglacial warming. But if this is so then increase in CO2 from 180-280ppmv should be responsible for temperature change of almost 10 deg K. And further increase from 280-380ppmv after the beginning of Industrial revolution should add antoher 10 deg with same sensitivity (or maybe little bit less to account for logharitimic nature of CO2-temperature relation). If we attribute all warming in the last 100 years to CO2 we obtain less than 1 deg, somewhat like 0.7 deg K, which is 10 times less than CO2 sensitivity derived from geological record Either Milankovitch related warming has some other, non CO2 very strong feedback, or in interglacial warming has some unknown and very, very strong negative feedback. Tertium non datur. And in both cases CO2 sensitivity is very low.
I don’t see in mainstream literature solution for this problem.
Re: Ivan (#146),
You underestimate Hansen. He has two sensitivities, short-term of 3K and long-term of 6K+. Cunning eh?
Oh, dang it, I forgot this:
http://en.wikipedia.org/wiki/Effect_of_sun_angle_on_climate
Thanks, MrPete. I have a fairly detailed discussion of the issue over at Watts Up With That for anyone who might be interested. I’m also looking for further resources regarding the total energy moved by a typical thunderstorm in all forms (latent, sensible, work, electrical, and all) if anyone has any.
All the best,
w.
Re: Willis Eschenbach (#163),
Willis:
Very nice, interesting post. You might want to ask Steve to run it as a guest post here, as the CA bunch is (I think) better-read in the climate literature than at WAWT. So you might get more useful feedbck here.
I presume you’ve read through Lindzen’s and Spencer’s thoughts on cloud feedback. Clouds do seem to be the “elephant in the living room” for climate-control feedback. It’s remarkable how little attention this topic has gotten from the professionals.
Thanks again for writing up your thoughts.
Best regards,
Pete Tillman
Re: Willis Eschenbach (#163),
At the risk of annoying our host, I’ll also post a link to Freeman Dyson’s doubts about AGW:
http://www.e360.yale.edu/content/feature.msp?id=2151
Required reading as a followup to his recent NY Times profile,
— which was agenda-driven and quite misleading. Dyson may not be an IPCC contributor, but he’s a very sharp scientist, and a cagey old bird, with an outstandingly sharp BS detector. And remarkably articulate. One of my scientific heroes.
Cheers,
Pete Tillman
MrPete : Are you sure the thunderstorm was not ahead of a cold front ? Then the cooling you felt was mainly due to the inflow of cooler air.
Re: Ulises (#164),
If it was, the climate didn’t know. I just googled a bit and came up with some interesting graphs.
The following links are to a citizen weather station not *too* far from where we were driving. Remember, we were heading west at ~60mph, so we passed through the storm rather quickly.
Here’s the graphs for that afternoon:
http://www.wunderground.com/weatherstation/WXDailyHistory.asp?ID=KCOMATHE2&month=6&day=23&year=2009
Notice how the temperature plummeted as the storm approached and hit. (In this case the measurement goes from 80 down to 60)…and then begins to rebound except it is evening and the sun is gone.
The next day shows a very different pattern.
Doesn’t prove much of anything, but it’s pretty intereting evidence that big storms cause a lot of heat to escape…
Ulises raises an interesting issue, that there are a couple kinds of thunderstorms. Some are thermally driven, and some are driven by mechanical elevation of warm air.
.
Thermallly driven thunderstorms are formed purely by local heat. These are found year-round in the tropics. They form over a local “hot-spot”, an area which is slightly warmer than its surroundings. This can be a warm patch of ocean, or a plowed field with grassland around it. They cool the surface through a host of phenomena, including:
.
1. Wind driven evaporative cooling. Once the thunderstorm starts, it creates its own wind around the base. This self-generated wind increases evaporation in several ways, particularly over the ocean.
.
a) Evaporation rises linearly with wind speed. At a typical squall wind speed of 10 metres/second (20 knots), evaporation is about ten times higher than at “calm” conditions (conventionally taken as 1 metre/second).
.
b) The wind increases evaporation by creating spray and foam, and by blowing water off of trees and leaves. These greatly increase the evaporative surface area, because the total surface area of the millions of droplets is evaporating as well as the actual surface itself.
.
c) To a lesser extent, surface area is also increased by wind-created waves (a wavy surface has larger evaporative area than a flat surface).
.
d) Wind created waves in turn greatly increase turbulence in the boundary layer. This increases evaporation by mixing dry air down to the surface and moist air upwards.
.
e) As spray rapidly warms to air temperature, which in the tropics is often warmer than ocean temperature, evaporation also rises above the sea surface evaporation rate.
.
2. Wind driven albedo increase. The white spray, foam, spindrift, changing angles of incidence, and white breaking wave tops greatly increase the albedo of the sea surface. This reduces the energy absorbed by the ocean. Geigers The Climate Near The Ground (my bible) puts the noon albedo of a calm sea surface at 2.1%, and a rough surface at 13.1%. This does not include the reflection or interception of light by foam, spindrift, and spray.
.
3. Cold rain and cold wind. As the moist air rises inside the thunderstorm’s heat pipe, water condenses and falls. Since the water is originating from condensing or freezing temperatures aloft, it cools the lower atmosphere it falls through, and it cools the surface when it hits. In addition, the falling rain entrains a cold wind. This cold wind blows radially outwards from the center of the falling rain, cooling the surrounding area. Outside the tropics ‘hydrometeors” (any form of falling water) from thunderstorms include snow, sleet, and the like.
.
As an aside, to me this is one of natures most ingenious tricks. In science, we know that the energy naturally only flows one way – from hot to cold. But consider the thunderstorm. It forms chunks of ice, and moves them from the cold atmosphere to the surface. It’s moving frozen objects to a hot surface in order to cool it down … to me, that’s a neat trick. It is a curious form of latent heat transfer. It cools the surface and continues cooling it until that latent heat debt is repaid. And from there, to re-enter the cloud cycle, all that melted water (at 0°C) has to be warmed up to evaporation temperatures. Another chunk of energy is needed to do that. Hydrometeors represent a huge energy transfer overall.
.
4. Albedo increase from increased reflective area. White fluffy cumulus clouds are not tall, so basically they only reflect from the tops. On the other hand, the vertical pipe of the thunderstorm reflects sunlight along its entire length. This means that thunderstorms shade an area of the surface out of proportion to their footprint, particularly in the late afternoon.
.
5. Modification of upper tropospheric ice crystal cloud amounts (Lindzen 2001, Spencer 2007) . These clouds form from the tiny ice particles that come out of the smokestack of the thunderstorm heat engines. It appears that the regulation of these clouds has a large effect, as they are thought to warm (through IR absorption) more than they cool (through reflection).
.
6. Enhanced night-time radiation. Unlike long-lived stratus clouds, thermally-driven cumulus and cumulonimbus generally die out and vanish as the night cools, leading to the typically clear skies at tropical dawn. This allows greatly increased nighttime surface radiative cooling to space.
.
7. Delivery of dry air to the surface. The air being sucked from the surface and lifted to altitude is counterbalanced by a descending flow of replacement air emitted from the top of the thunderstorm. This descending air has had the majority of the water vapor stripped out of it inside the thunderstorm, so it is relatively dry. The dryer the air, the more moisture it can pick up for the next trip to the sky. This increases the evaporative cooling of the surface.
.
So that’s the first kind, thermally driven.
.
The other kind, the frontal kind you describe, arises when a cold wedge of air is pushing under a warmer air mass. In that case, the surface cools immediately as you point out. Often that happens in advance of the actual clouds forming. In this case, they are not driven by heat from underneath. The wedge of cold air forces the warm air upwards. At some point moisture condenses, and as the warm air mass continues to ascend, clouds and thunderstorms form.
.
You are correct that a quick shift in temperature can also occur from a frontal storm. MrPete’s maps are conclusive in that regard, there’s no room for doubt. No sign of frontal passage in the barometric record, usual down at dawn, up at dusk. Winds drifting around the compass (the noon jump is from crossing the edge of the graph). Rainfall in the evening. My reading is local, thermally-driven thunderstorm. By the late timing of the rain, I would guess it was triggered by the late afternoon cooling of the atmosphere. This can encourage condensation.
.
My point is slightly different, however. It is that all thunderstorms reduce the surface temperature. Thunderstorms are solar-driven machines that provide us surface dwellers with ice, chilled water, and air conditioning. The planet receives enough energy to make it unbearably hot. Thunderstorms ice down and cool down and air condition the surface, preventing the planet from overheating.
.
And this cooling and air conditioning is a result of all thunderstorms, whether they are of frontal or thermal origin.
.
Not only that, but thunderstorms form spontaneously in response to rising temperatures. They don’t care what’s making the surface a bit hot, whether it’s CO2 or plowed fields or both. When a certain threshold is passed, they spring into existence and immediately start putting out the fire, cooling the hot surface and reflecting away the sun.
.
As a some-times inventor myself, I can only stand in awe of the intricacy of the thunderstorm’s engineering. A solar powered air conditioning machine that comes on automatically when it is needed, with chilled water and blasts of cool air delivered to my front door. An incredible machine.
.
I want the same thing for dirt. When a part of my carpet is too dirty, I want a little spinning tornado to form spontaneously, wander around my carpet, whisk away the dirt, and then disappear …
.
Ulises, interesting question. MrPete, best support of a weather claim I’ve ever seen, case dismissed. Everyone, my best wishes.
.
w.
willis, a nice paper is Fierro Simpson et al 2009 . Its reference list is a compilation of articles on tropical convection and precipitation, with some commenting on the energy aspects.
Regarding that paper, it discusses the air entering tropical thunderstorms from the boundary layer. The air entering the thunderstorm is described as a mix of “dry” (lower-humidity) air parcels and “wet” (higher-humidity) air parcels, not just wet ones.
The ratio of the parcels, and their moisture content, affect, I believe, the ultimate buoyancy of the mix and thus its ultimate altitude and temperature (the higher the colder). This ultimate temperature determines the dryness of the mix (the higher the colder the drier).
I wonder if, in a warmer world, those “dry” air parcels have a higher moisture content. The mix entering a warmer-world thunderstorm then has more moisture and buoyancy and reaches a higher, colder, drier destination in the tropical upper troposphere. This helps create a drier tropical upper troposphere, which helps remove heat from earth, sort of a negative feedback in this warmer-world.
David, many thanks for the fascinating cite. I’ll need time to digest it, but it looks like it’s full of good stuff.
w.
4 Trackbacks
[…] more: Cloud Super-Parameterization and Low Climate Sensitivity 11 Jun 09 | […]
[…] Cloud Super-Parameterization and Low Climate Sensitivity by Steve McIntyre on June 11th, 2009 […]
[…] post on Watts Up With That introduces and enhances a recent post on Climate Audit describing strong negative cloud feedbacks found by the Climate Process Team on […]
[…] the IPCC projections — like this new MIT study. Recent discussions of water vapor and cloud cover feedbacks at ClimateAudit are fascinating, and illustrate both the political urgency surrounding […]